SilverPlatter

  • ERL
  • WEBSPIRS
  • WINSPIRS 4.01
  • MACSPIRS
  • Z39.50
  • SILVERLINKER
  • WEBADMIN
  • HCU
  • QUIKDATA
  • DATABASES

ERL

Statistics

General

Database Instillation

Configuration Options

Operating System Configuration

Error Messages

Installation

Will I be able to save my ERL v2.x statistics in moving to ERL v4.x?
Before you install ERL v4.0 we recommend that you perform a system backup and export any statistics you wish to save.

What files contain the statistics information in ERL v4?
The statistics files are stored in the /sproot/erlstat directory and are named 001.dat and 001.idx.

What are the functions of the stattemp and statkeeptemporaries parameters and should I change these?
The stattemp and statkeeptemporaries parameters may be set in the erlsrvr.cfg file in /sproot/erlstat. If you set statkeeptemporaries=1 (default is statkeeptemporaries=0), ERL will save the temporary stat files in the directory specified in the stattemp parameter. You would normally only want to save these temporary files if you were diagnosing a problem.

I have just installed ERL and now I want to create user accounts. How do I log in with the administrative client if I haven't yet set up any usernames? The default administrative user in new ERL installations is admin/admin. You can log in with this client to administer the server. For security reasons, it is recommended you change the administrative password after initially logging in to something other than the default. You must have WebAdmin or ERLAdmin installed to create user accounts on your ERL server.

Before you can login for the first time you will also need to ensure your IP address is authorized on the ERL Dbserver by editing the ipincl file in /sproot/cfiles.

What has changed in Dbserver 4.0 as compared to Dbserver 2.x?
* dbmenu.cfg now in /sproot/cfiles
* by default the NOAUTHLIST is set to NO for all new databases
* /sproot/stats is removed
* pstats.dat/idx and dstats.dat/idx are removed
* /sproot/erlstat now stores 001.dat/idx as the new statistical files
On NT, ERL Dbserver is no longer started as "System Account", but as NT user "ERL".

In ERL 4.0, what is "readmenu" used for?
It causes dblbuild to re–read the dbmenu.cfg file and to rebuild dbl.txt and dbl.cfg. This script will allow you to rebuild the SPIRS client database selection screen (dbmenu.cfg).

The file is located in the /sproot/bin/command directory. You can run this at the command prompt or console depending on your operating system (NT or UNIX). If you have problems running the "readmenu" script from /sproot/bin/command directory, you may copy "readmenu" to the /sproot/bin directory for better results. Then run it and it will update the dbmenu.cfg file without the need to restart the ERL server.

What are 'Lister' and 'Reader' and how do they function?
Lister and Reader are utilities located in /sproot/bin that provide legible versions of statistic information. You would run lister on a Unix platform as follows:
cd /sproot/bin

./lister ..erlstat/001

Reader enables statistic reporting with a variety of keys. First, you must create a stattemp directory within the sproot directory. From the /sproot directory, type:

Bin/reader uname=a* dbid=*

(See full list of keys below)

Reader will create a file in the stattemp directory. In order to read this file, you must run ctexport as follows:

/sproot/bin/ctexport –v filename.dat >filename.txt

Use More filename.txt to read the content of the statistics information.

A full list of keys and their meanings are given below. Note that abbreviations used by Reader are not the same as BrlAdmin or WebAdmin

Key Description
Fam Database family, (set or group), the 2 letter id. (KEY)
Dbid Database id. (KEY)
Pfolio Portfolio Name (KEY)
Ip IP address key (KEY)
Uname Username (KEY)
Detail TOTAL, DAILY or HOURLY
Filter EQL, GTE, SEQL or SGTE
Uid User id (Numeric) (KEY)
Hist Set to 1 if the files are old and no dates are given to finish processing at the last entry in the file, rather than the current time. No need to use this normally.
Statpath An alternative path for the statistics files, eg d:\sproot\special\erlstat
Export Set to 0 to avoid calling ctexport
Session Session key. (KEY)
Start Date to start the report. Format is [dd/mm/yy[yy]][:hh[:mm[:ss]]]. There must be NO spaces. Omitted items default to NOW.
End Date to end the report. Format as above.

Items marked (KEY) are statistics keys. In all cases except userid (which is mainly used by ERLAdmin) the key may be a * or any prefix followed by a star. For IP addresses, this is true, but the full story is more complex. For user id, allowed values are a valid (numeric) userid, or a *.

Solaris CD not mounted, but won't eject from the drive (Solaris).
After typing "mount" or "/etc/mount" to make sure the drive is not showing in the list of mounted volumes, try typing "eject" at the prompt and pressing the return key. This should eject the CD from the drive.

Use truss to diagnose the status of a process (Solaris).
The truss command executes the specified command and produces a trace of the system calls it performs, the signals it receives, and the machine faults it incurs. Each line of the trace output reports either the fault or signal name or the system call name with its arguments and return value(s).

Truss is run as user root.

Examples:

truss ./dbserver
truss ./dbinstall
truss ./dblbuild


Truss is terminated using Ctrl–C

What are the current SilverLinker Publisher URLs?
As of March 2000 they are:

Publisher url_prefixes

Academic Press http://www.idealibrary.com/
Academic Press http://www.europe.idealibrary.com/
CatchWord http://www.catchword.com/
DeGruyter http://www.degruyter.com/
Ebsco http://www.ebsco.com/
IEE http://ioj.iee.org.uk/
Information Quest http://www.eiq.com/
Karger http://www.online.karger.com/
Kluwer Academic Publishers http://kapis.www.wkap.nl/
NRC Research Press http://www.nrc.ca/cisti/
Project Muse http://muse.jhu.edu/
Science Direct http://www.sciencedirect.com/
SIAM http://epubs.siam.org/
Springer http://link.springer.de/
Springer http://link.springer–ny.com/
SwetsNet http://www.swetsnet.nl/
SwetsNet http://www.swetsnet.com/

There are no SilverLinker links appearing in my database. What do I need to check?
1) Verify that the database you are using works in combination with SilverLinker. Check the following URL for the list of enabled databases: http://www.ovid.com/site/products/tools/silverplatter/silverlinker_enabled.jsp

2) Check the setting in the /sproot/bin/erlsrvr.cfg file and make sure that it does not have a remark statement (/*) in front of the following: provider.SLINK=perlprov

3) Check that you have the "FTXT" field displayed in your client. If it is not, it may be enabled as a default field for all databases using WebSPIRS 4 by entering FTXT in the /webspirs/bin/webspirs_user.cfg file in the line "default.fields.p=".

4) Verify that SilverLinker is licensed, mounted and enabled through ERL dbserver user interface.

5) Check through ERLAdmin or WebAdmin and verify that SilverLinker is authorized.

6) Possibly the search(es) you are testing are ones which do not have associated links. A sample search for Medline ( any 1997 coverage) is "j–surg–res".

What is a Database ID?
A database ID is a 4 or 8 character unique identifier for a database.
For instance:
– AE16 is the database ID for: EconLit 1969–8/98
– CCOC0123 is the database ID for: Current Contents 7 Eds. 1998/10/10–1998/10/16 weekly update.

What is a Volume ID?
A Volume ID identifies a SilverPlatter CD.

– In most cases there is only 1 database on a CD. In that case the Volume ID is the same as the Database ID.
– In some cases 1 CD can contain multiple databases. In that case there is 1 Volume ID for multiple Database ID's
– It is also possible that 1 database consists of multiple volumes. In that case there is 1 Database ID that has multiple volume ID's

THE "MESSAGE OF THE DAY" FUNCTION IN ERL V4 AND ERL V2
ERL v4 includes a Message of the Day (MOTD) function. As the ERL administrator, you can make a message of the day available to all user accounts that login to the server.

Create a text file called "motd" that includes the text of your message. You can use the file /sproot/masters/motd.eg as an example. The location of this file on the ERL server, is in the directory /sproot/cfiles. In order to create this file you are advised to use Webadmin 4.0 or use a editor like "vi" for UNIX platforms or Notepad for NT.

In case you only have to display the "motd" to WebSPIRS users, you may want to enrich the message by writing it in html rather then plain ASCII

You can also place a motd file in the directory /sproot/cfiles/userid where userid is the internal ID associated with a user account (found in the "Ident" field on the View/Edit User screen of the ERL Admin client). This message will go to the specific ERL user account associated with the directory.

If you have multiple servers, you can create a message file on each server. Clients will receive a message from each server that they access.

Please note that while the Message of the Day function has been implemented in the ERL v4 server software, it must also be implemented in the client software. In order for users to see the message of the day, they must use the following versions of client software:

WinSPIRS v2.1 or higher
MacSPIRS v2.4 or higher
WebSPIRS v3.1 or higher

Typing mistakes can cause improper processing of ipincl/ipexcl
The ipincl and ipexcl files in /sproot/cfiles are used to control access to the ERL server per IP addres. Certain typing mistakes such as embedded blanks in a IP address will prevent the IP addess checking mechanism from working correctly. As a possible consequence ALL IP addresses would be allowed/denied accessing the ERL server.

Example: 193.171.95.(blank)0(blank)255.255.255.0 instead of 193.171.95.0(blank)255.255.255.0

If the first blank is removed the ipincl file will be properly processed by the ERL server.

How do I find the current ERL version Number on my server.
Start the ERL user interface (dbserver menu), the first line will display the ERL version number.

More on ERL ctree files and power failures
If due to a power failure the ERL server goes down, there is a chance that one or more of the so called "ctree" files become corrupted. The ctree files store information about users, licenses, statistics and installed databases. It is important to make a regular back–up of your /sproot directory regularly as it is not always possible to recover ctree files that become corrupt.

If you are interested on how to check the ctree files on the ERL server we have the following information:
The ctree files are always kept in pairs (.dat and .idx extensions).
For dbserver 2 and 4: in /sproot/files, dbinfo.dat/idx, users.dat/idx, pfolio.dat/idx, authlist.dat/idx
For dbserver 2 only: in /sproot/files, pstats.dat/idx and dstats.dat/idx
For dbserver4.04 only: in /sproot/erlstat, 001.dat/idx.

To check for corruptions in the ctree files, you should run the "ctexport" program on the ctree file with the .dat extension. When corrupted the ctexport will abort with an error. Make sure always to run as the erl user.

Example:
– cd /sproot/files
– ../bin/ctexport dbinfo.dat

If the file is OK you will see an ascii representation of the dbinfo file on your screen it would be best to redirect the output to a file, if the ctree files are large (../bin/ctexport dbinfo.dat > dbinfo.txt). Use ctexport –v to obtain field information.

Can I use SilverPlatter Software with Thin Client Solutions (NT Terminal Server and Metaframe, Citrix Winframe)?
Thin Client Solution (NT Terminal Server and Metaframe, Citrix Winframe) are not currently supported operating systems for SilverPlatter software.

However:

1. As long as you run a supported browser on theThin Client Solution Terminal and from there point to ERL dbserver v.4 (i.e. WebSPIRS v.4), you will be able to offer the full ERL dbserver v.4 functionality.

2. Several customers and partners of SilverPlatter are using SilverPlatter software (WebSPIRS 4 and dbserver 4) installed on Winframe or NT terminal server (+metaframe) and do not experience problems.

3. Having WebSPIRS or ERL dbserver running on the NT terminal server puts an extra load on the machine that will have a detrimental impact on the system performance.

4. WinSPIRS (any version) can not be installed on NT terminal server.

Medline Express 2000 edition for ERL
Medline Express 2000 edition consists 5 sets of 3 disks containing the archives from 1966–1999, and 1 update disk that is replaced each month covering January 2000–present. Each archival set that consists of 3 CD volumes with 3 Volume ID's is recognised by ERL as a single database with 1 database ID. An overview of how the 2000 edition is build up can be found in this Tech Tip.

When you have 1 CD drive specified as input device in /sproot/files/devices, you will have to start with the first disk of a set. ERL will recognise this as a database and will then FIRST ask for the other 2 volumes to check if you have the complete set available. Then it asks for the first disk again, and only then it will start copying files. If you choose for CD level install, it will only copy the *.DIR file.

If you have 2 CD drives specified as input device, you are advised to edit the devices file prior to installation of Medline Express making sure ERL uses only 1 drive.

If you have 3 drives or more available and specified in the devices file, then ERL will find the database and install it in one go.

Please note that in case you do not have a license installed, that allows you to copy to hard disk, it will default to CD–level install and only copy the .DIR file. For CD level install you do need to have sufficient CD drives. For instance if three CD volumes make up one Database, then you need at least 3 CD drives permanently available for this database.

SP–071–xxx Disc 1, 2000–present VolID MX41 DbID MX41 Up to 690,000,000 bytes
SP–071–241 Disc 2, 1997–1999 VolID MXAA DbID MXAD
SP–071–242 Disc 3, 1997–1999 VolID MXAB
SP–071–243 Disc 4, 1997–1999 VolID MXAC 1,803,590,232 bytes
SP–071–244 Disc 5, 1994–1996 VolID MXAE DbID MXAH
SP–071–245 Disc 6, 1994–1996 VolID MXAF
SP–071–246 Disc 7, 1994–1996 VolID MXAG 1,790,855,024 bytes
SP–071–247 Disc 8, 1991–1993 VolID MXAI DbID MXAL
SP–071–248 Disc 9, 1991–1993 VolID MXAJ
SP–071–249 Disc 10,1991–1993 VolID MXAK 1,571,973,269 bytes
SP–071–250 Disc 11,1986–1990 VolID MXAM DbID MXAP
SP–071–251 Disc 12,1986–1990 VolID MXAN
SP–071–252 Disc 13,1986–1990 VolID MXAO 2,052,981,400 bytes
SP–071–250 Disc 11,1980–1985 VolID MXAQ DbID MXAT
SP–071–251 Disc 12,1980–1985 VolID MXAR
SP–071–252 Disc 13,1980–1985 VolID MXAS 1,654,570,753 bytes
SP–071–256 Disc 17,1966–1979 VolID MXAU DbID MXAZ
SP–071–257 Disc 18,1966–1979 VolID MXAV
SP–071–258 Disc 19,1966–1979 VolID MXAW 1,991,455,431 bytes

Databases are not appearing in the list of available databases, even though it is installed, mounted, and enabled. ERL v4
When a license is initially installed in ERL version 4, the default for the "Use Authorizations" value is Yes. This means that users must be specifically authorized to that license portfolio. For sites that do not authorize users specifically to portfolios, this setting will prevent the database from displaying in the list of available databases.

Solution: Change the User Authorization value from Yes to No. In WebAdmin deselect the "UseAuths" checkbox in the Databases section. In ERLAdmin, switch the Noauth value under Databases/View–Edit Database from N to Y.

Error: bad file number
This error may appear when attempting to install a database on your ERL 4.0x server. The message indicates a permissions problem on the destination directory.

To correct this error:
1. make note of the destination location– for example: /sproot/nodes/disk2
2. ensure that the permissions on the disk2 directory are drw–r––r––
– to change the ermissions, enter: chmod 644 disk2
3. ensure that the owner and group are set to erl
– to change the owner, enter: chown erl disk2
– to change the group, enter: chgrp erl disk2

What is a Database ID?
A database ID is a 4 or 8 character unique identifier for a database.
For instance:
– AE16 is the database ID for: EconLit 1969–8/98
– CCOC0123 is the database ID for: Current Contents 7 Eds. 1998/10/10–1998/10/16 weekly update.

What is a Volume ID?
A Volume ID identifies a SilverPlatter CD.
– In most cases there is only 1 database on a CD. In that case the Volume ID is the same as the Database ID.
– In some cases 1 CD can contain multiple databases. In that case there is 1 Volume ID for multiple Database ID's
– It is also possible that 1 database consists of multiple volumes. In that case there is 1 Database ID that has multiple volume ID's

On most SilverPlatter CD's the Volume ID is printed on the CD above the "SP–xxx–xx" number.

What is a Database family ID or Set–ID
The Database family ID, also known as Set–ID, are the first 2 or 4 characters from a database ID.
This code identifies a group of databases that belong to the same database family. The Database family ID of the databases
you are subscribing are listed on the database license in the line for "Databases".

For instance:
– The Family ID of Econlit is: AE
– The Family ID of Medline Express is: MX
– The Family ID of Current Contents 7 Editions weekly updates is: CCOC
– The Family ID of Current Contents 7 Editions archives and cumulative updates is: CCON

Should I install a database license before installing my databases
Yes, we do recommend you install the database license before you start the installation of the databases that are covered by your database license. The values in your database license are needed in order for the ERL dbserver to perform a "magnetic copy install". If you have not installed the database license, or if the value for "allow magnetic copy install" in your license = No, then ERL dbserver will not be able to install the database using the magnetic copy install option from the dbserver menu..

Installing a database on CD rom level (no magnetic copy), or installing a database as a magnetic link install (database copied to hard disk manually) can be done without installing a license first. However without a valid database license, these databases can not be accessed by any SPIRS client.

What exactly is an "EXPRESS" database
In most cases 1 SilverPlatter CD contains 1 SilverPlatter database for a certain date range. Therefore a subscription to a certain database product that covers many years, can consist of multiple CD–roms.

Some of the larger databases (Medline Express, Inspec and Embase Express) that are provided by SilverPlatter on CD–rom, contain so called multi volume databases. This means that 1 database, with 1 database ID consists of multiple CD volumes each with a volume ID (similar to a database ID). ERL dbserver wants to install these multiple volume databases as one database, available under one database ID. This saves a lot of disk space, and therefore SilverPlatter is able to fit the EXPRESS databases on a smaller number of CD–roms.

How do I install an 'EXPRESS" database on ERL
It is important to realize that when you install the Multi Volume parts of the EXPRESS databases, the install procedure is slightly different from when installing a Single Volume database.

If you have specified a sufficient number of CD–drives in the /sproot/devices file, you can simply insert all the CD's of a set in to the drives and ERL will see the database and you can continue with the installation as normal.

In case you only have 1 CD drive available for a multi volume database, then it is important to always start with the first CD of a set. This is the CD with the lowest disk? number of the set. If you are unsure which disk is the first CD of a set, then please be advised to start with the CD that contains the file with the .DIR extension. The install process will not start copying the database to hard disk, but first ask for all the other CD's of the set to ensure they are available. Then the install process will ask for the first CD again and start the copy process.

Please note that the install process will request the CD's by their volume ID's and not by the disk number. Most CD's will have the volume ID printed on the CD above the SP number.

Please note that the install process will request the CD's by their volume ID's and not by the disk number. Most CD's will have the volume ID printed on the CD above the SP number.

Can I install an EXPRESS database as magnetic link
Yes, this is possible. A magnetic link install can be used to install databases that are already located on hard disk. This can be on the ERL server or a shared location on the network. How to install a database as magnetic link is explained in the ERL administrators manual page 4–5. The manual can be found at:
http://www.ovid.com/site/support/system_admin_manuals.jsp

Warning!: At any time when you want to do a magnetic link install, you have to ensure that the location of the database files is NOT the same as the location where ERL is supposed to write database files. This will prevent ERL trying to write a file to the same location as its origin. This situation would lead to an install failure and the loss of the *.DIR file or the DISC.VFF and *.DFF files.

To install an EXPRESS database as magnetic link, you can use the following additional instructions. In this example we take the Medline Express database MX9W that consists of three volumes MX9J,MX9K and MX9L:

1. create new subdirectories at a location where ERL does not write. For example on UNIX: /sproot/nodes/disk?/medline/MX9J, MX9K andMX9L. On NT for example c:\medline\MX9J, MX9K, MX9L.

2. copy manually all files from the CD volumes in to the corresponding directories on the ERL server

3. on UNIX platforms, ensure that all filenames are in Upper case after copying (else rename them). Solaris Specific Instructions :to mount the CD volumes manually on Solaris, please use the following mount command: mount –F hsfs –r –o nomaplcase . The nomaplcase option will mount the CD–ROM with all of the filenames and directory names presented in uppercase characters.

4. compare the filesizes from the CD rom with the files that you copied, before you procede. They have to match!

5. On UNIX and NT define DBIPATH (export DBIPATH= ) for the erldata directory. For UNIX platforms this is described in the manual on page 4–5. For NT you can use the following command: set DBIPATH=.

6. The pathname in this example for MX9W would be DBIPATH=/sproot/nodes/disk?/medline

7. Please note that you enter the DBIPATH name from the command line and this environment setting will only be valid for that specific console session.

8. Then install the database using the "search for candidates" and "magnetic link" install option as described in the manual.

Can I use DBIPATH on NT?
Yes, you can. For NT you can use the following command: set DBIPATH= Further instructions can be found in the ERL administrators manual on page 4–5. Please note that you enter the DBIPATH name from the command line and this environment setting will only be valid for that specific console session.

How does dbserver decide where to write the databases?
Regardless the type of database installation that you try to do (CD–rom level, magnetic copy, magnetic link or upgrade installation), the ERL database installation process will always write at least one of the database files to hard disk. The ERL dbserver will be in control of where it will write the database files.

On NT: ERL will find the location where it is allowed to write in the /sproot/files/nodes file. If there are multiple locations listed in the nodes file, there is an algorithm in the ERL install process that will decide which location will be used. In general this algorithm will always try to fill the location with the smallest available free disk space first. If you want to prevent ERL from continuing to write to any of the locations listed in the nodes file, you simply take the location out of the nodes file. This does not affect the databases already installed in that location.

On UNIX: ERL will look for any directory available under /sproot/nodes. If there are multiple directories in /sproots/nodes, then there is an algorithm in the ERL install process that will decide which location will be used. In general this algorithm will always try to fill the location with the smallest available free disk space first. If you want to prevent ERL from continuing to write to any of the directories under /sproot/nodes, you simply can make this directory write protected for the erl user. As long as the erl user can still read from this directory the databases that were already installed are not affected.

Can I copy databases to hard disk manually, so I keep control over the location of my databases?
Yes you can, as long as you do not copy the databases in to a location were ERL normally writes the database information. (i.e the locations listed in the /sproot/files/nodes file for NT or /sproot/nodes/subdir for the UNIX platforms) After you copied the databases to hard disk, you can install that using the "magnetic link install" option. For the UNIX platforms please note that after copying the database files to hard disk, the file names MUST be in Upper Case notati

Can I prevent dbserver writing on to certain filesystem under /sproot/nodes
On UNIX: ERL will look for any directory available under /sproot/nodes. If there are multiple ldirectories in /sproots/nodes, then there is an algorithm in the ERL install process that will decide which location will be used. In general this algorithm will always try to fill the location with the smallest available free disk space first. If you want to prevent ERL from continuing to write to any of the directories under /sproot/nodes, you simply can make this directory write protected for the erl user. As long as the erll user can still read from this directory the databases that were already installed are not affected.

What to do when a database installation does not complete successfully
In this example we look at a situation where the database MX29 has failed to install because the server had an unexpected "power down" during the database installation.

MX29 is a single volume database. Please note that the same rules are valid for failed installations of multi–volume databases. If you feel insecure about the using the instructions below, please do not use them and contact SilverPlatter Technical support directly instead.

A failed database installation can make it impossible to re–install the database or it can even block the dbinstall process (option 6. from the main dbserver menu) from starting. Another side effect can be that the dblbuild process is no longer able to create an updated database list. Such a situation, is always caused by inconsistencies between the following three directories: /sproot/spdbvols, /sproot/install/dbdata and /sproot/install/voldata

1. make sure you first unmount and remove the database in question using the dbserver menu.
2. remove the directory that is called after MX29 from /sproot/spdbvols.

*****
NB! If it is just dbinstall not finding the database, it can be that this is all you need to do, then step 3–11 are redundant. If not continue with 3.
*****

3. SHUTDOWN dbserver
4. remove the MX29 file from /sproot/install/dbdata
5. edit the db.map file in /sproot/install/dbdata to remove the entry for MX29. Please note there is a checksum on the first line that reflect the total number of installed databases. Therefore it has to be decreased by 1.
6. remove the MX29 files from /sproot/install/voldata including the MX29.VFR file
7. edit the vol.map file in /sproot/install/voldata to remove the entry for MX29 (3 lines!). Study the format of this file carefully before you remove the lines for MX29! You'll notice that multi–volume databases are in a special format Please note there is again a checksum on the first line that reflect the total number of installed databases that has to match the checksum in the db.map file. Please also note the format and counting of a database like MX9W that consists of MX9J, MX9K and MX9L.
8. it is a good idea to check if all information in /sproot/spdbvols, /sproot/install/voldata and /sproot/install/dbdata are consistent with each other.
9. clear /sproot/nodes/disk?/MX29 to remove the residuals of MX29. (For NT this will the location(s) specified in the \sproot\files\nodes file
10. start dbserver. At this stage MX29 is removed but might still be listed as a possible candidate to mount. This is because dbinfo.dat still expects that database on hard disk. However this does not affect ERL
11. MX29 can be installed normally again. We advise you do this even if you no longer need this database. If you do not install MX29 again, there might be a residual left, because the dbinfo.dat (a binary file you can not edit) still remembers MX29 as a installed database. Therefore MX29 will remain in the list of databases that can be mounted. This is not harmful but may be annoying as it will show up every time. You are advised to reinstall the database and remove it properly to prevent this problem

If you feel insecure about anything mentioned above please do not use use these instructions. SilverPlatter TechnicaI Support can remote login to your ERL server using telnet and make the edits for you, or alternatively you could send us a listing ("ls –l" or "dir") of the directories mentioned above including the actual vol.map and db.map file. Then we can edit the vol.map and db.map files and send them back to you with specific instructions on how to proceed.

After I installed a database it does not show up in the database list of my SPIRS client.
This can have many reasons. Below you will find a checklist you can use to determine why.

The database list that is presented to the SPIRS ERL client is a representation of the dbl.cfg file located in /sproot/files. When the user is logging in to the ERL server the information from the dbl.cfg file plus the structure from the /sproot/files/dbmenu.cfg file is used to present a database list to the SPIRS client. The dbl.cfg file is a human readable ASCII file that should not be changed using an editor. The dbl.cfg file is renewed each time the dbserver is started or when unmounting/mounting a database with the dbserver menu. The actual process that performs this task is the dblbuild process.

1. Is there a valid "master" license installed for this database. To check, use option "5. License administration" from the main menu, and then use option "4. Display license" to display the master license for the database family of the database that does not show up. The master license is the license that starts with L–. If the master license has expired or if the database family ID of the database is not covered by the license then the database (although installed correctly) will NOT show in the database list on any SPIRS client. However, it will show up in the dbserver menu.

2. If the there is a valid master license, but the user logging in to the ERL server is only authorised for a sub–portfolio of the master license for this database family, then use erladmin or Webadmin to verify if this sub–portfolio is still valid and covered by the settings in the sub–portfolio.

3. Is the user that is logging in to the server authorised for this databasese? Use erladmin or Webadmin to verify this. The "admin" ERL user or ERL users with admin equivalent privileges, will be able to see all databases without the need for authorisation.

4. Is the database "mounted"?. To check, use option 1. Show mounted database(s)" from the dbserver menu.

5. Is the database enabled?. To check use option 4. Enable/disable database(s) from the dbserver menu.

Before you continue with 6–11, you are advised to check your /sproot/bin/dbserver.log file for indications about the source of the problem.

6. Check if there is a dbl.cfg file in /sproot/bin. If so, remove this file and unmount/mount a database to update the dbl.cfg file in /sproot/files.

7. Check if the database that does not show up in the database list on the SPIRS client is actually listed in the dbl.cfg file. If it is listed in the dbl.cfg file then go back to points 1–5 to find out what is causing the problem.

8. If the database is not listed in in dbl.cfg, then remember the date and time of creation of the dbl.cfg file and unmount/mount a random database using the dbserver menu. Then check if the dbl.cfg file was re–created by comparing the (new) date and time of creation. If the dbl.cfg file was recreated but still does not contain the missing database then continue with 10.

9. If the dbl.cfg file was not recreated, this indicates the the dblbuild has died or became unstable. In that case stop the dbserver and start it again. On NT stop/start the ERL service, on UNIX use the dbserver menu to shutdown, and start the server from /sproot/bin with the ../scripts/start_server command. Check if the dbl.cfg file was updated after starting the server and if the problem database is now included in dbl.cfg.

10. If the dbl.cfg file was recreated but still does not contain the problem database / if the dblbuild process does not want to start/ If the dbl.cfg file still was not recreated this indicates an inconsistency in the /sproot filesystem. First verify if the dbinstall process will start by choosing option "6. Database installation" from the main dbserver menu. If it is not possible possible to enter this menu, then you are advised to solve this problem first. If dbinstall does start and you can enter the database installation menu, then continue with 11 or try to remove and re–install the problem database again.

11. Check and compare the following directories for inconsistencies and be advised to start with those databases that were last installed:
– /sproot/install/vol.map should be consistent with the contents of /sproot/spdbvols. i.e. for each volume listed in /sproot/install/voldata there should be a directory in /sproot/spdbvols.
– Each subdirectory of /sproot/spdbvols represent a database volume and contains pointers to the location of the actual database files. These pointer location can be listed with "ls –l" on UNIX platforms and on NT you must use the type command followed by the file name. Please note that the pointer to a file is build up out of two parts. A general part called after the volume ID, and a second part called after the file name.
– Check if all the pointers are pointing to a database file. If pointers are missing or pointing to non–existing or unreadable files, then you have found the source of the problem and are advised to remove and re–install this database

What to do when ERL suggests to remove a "archive" database when installing new updates.
When you are installing new updates for your databases on ERL and older updates are still installed, then ERL will suggest to remove these older updates as they overlap with the new updates. However for some databases this suggestion is not always correct.

The numbering scheme of The SilverPlatter database ID's determines if a database is an update or an archive database:
– Each database ID that is ending at a letter is supposed to be an archive database.
– Each database ID that is ending at a number is supposed to be a database update.

The algorithm in ERL that suggest to remove older updates is very simple. It only looks at those database ID's that end with a number. If it then can find an older database that ens at a lower number from the same database family, it will suggest to remove the older database. Weekly database updates in our current numbering scheme are all ending at a letter and will be ignored by ERL when installing new database updates. Unfortunately a few archive databases from a couple database families do end at a number. For instance IB12 which is the database ID for the INSPEC 1996 archive database. This can actually lead to a situation where ERL will prompt the administrator with a suggestion to remove IB12 when installing the monthlyupdate database for 1999. In this case this is not correct and the message should be ignored.

After I "untarred" databases from tape or downloaded them with QUIKdata, how do I continue.
After you "untarred" or downloaded the databases from tape or with QUIKdata to you hard disk, you can perform a magnetic link install or a magnetic copy install. Be aware that when you perform a magnetic copy install you actually duplicate the database on your hard disk. To save disk space you best remove the database from the location it was "untarred" or downloaded to, after you successfully installed it with "magnetic copy install". When you later remove this database with option "2. Remove database(s)" from the dbserver menu, it will also remove the actual database files.

When you use "magnetic link install" this is not the case. The data stays in the same location as it was "untarred" or downloaded to and ERL will create the pointers in the /sproot/spdbvols directory for this database to point to the correct location. When you remove a database that was installed with the "magnetic link" option, it will not remove the database file, but only the pointers.

The location you have used to "untar" or download the database to is free to choose with the exception of the location that are in use by ERL to write databases to. That location can be on the ERL server but also a shared location on the network. How to install a database as magnetic link is explained in the ERL administrators manual on page 4–5.

Warning!: At any time when you want to do a magnetic link install, you have to ensure that the location of the database files is NOT the same as the location where ERL is supposed to write database files. This will prevent ERL trying to write a file to the same location as its origin. This situation would lead to an install failure and the loss of the *.DIR file or the DISC.VFF and *.DFF files.

ERL can't find the database as a candidate for installation.
This usually indicates that a previous attempt to install the database failed, and now the existence of an empty directory in /sproot/spdbvols called after the volume id of that database blocks the possibility to install that database.

ERL does not release the CD–drive after installation
This sometimes happens on Solaris or AIX servers. And most often with the Multi volume parts of the "EXPRESS" databases. Best is then to login as root and run the "umount" command on the device or mountpoint that is occupied but no longer needed by ERL.

How does the verification mechanism work?
At database installation time ERL keeps track of all the files that are copied to hard disks and all the pointer information that is created in /sproot/spdbvols. It will store details about file names, size and creation date and time for each database volume in a file with the extension .VFR in /sproot/voldata. This information is not a reflection of what is on the CD, but only of the database as it was installed.

If the database was not installed correctly the verification mechanism in ERL (option 5. from the database install menu) will not be able to repair the database installation. Only if the database was installed and used successfully, and then got damaged the verification mechanism can be used to repair the database. The verification mechanism wil be able to repair any of the pointer information or damaged database files. When it needs to get the original file back from the CD, it will automatically prompt for the correct CD.

There is a database that shows up in my list of mountable databases, and I can't mount nor remove that database. This is a residual from a database that was removed from the ERL server manually. The only way to remove this entry is by re–installing this database first, and then remove it correctly using the dbserver menu.

Can not intitialise disk manager message when trying to go in the database installation menu.
The first thing to do when you get this error ist to check for files in the /sproot/nodes/ directory. If any file exists in this directory, move them to another directory outside of /sproot/nodes. There should only be directories in /sproot/nodes!

A failed database installation can also block the dbinstall process (option 6. from the main dbserver menu) from starting. When option 6. Database installation is chosen from the dbserver menu, you actually start a new process on the server called "dbinstall". During the initialisation phase of this process several checks are performed. When dbinstall fails in the initialisation, it will prompt an error like "Can not intitialise disk manager". This situation, is usualy caused by inconsistencies between the following three directories: /sproot/spdbvols, /sproot/install/dbdata and /sproot/install/voldata.

To recover from this situation can be quite a puzzle. However in most circumstances the ERL administrator will have a pretty clear idea which database entries are causing the problem and when analysing the above–mentioned directories, it will be relatively easy to identify and repair the problem.

Make listings from all files in /sproot/spdbvols, /sproot/install/dbdata and voldata, and then a printout of the db.map and the vol.map files. Check and compare using the following instructions for inconsistencies and be advised to pay specific attention to those databases that were last installed or that are suspected to have caused the problem.

– /sproot/install/db.map should be consistent with the contents of /sproot/install/dbdata. I.e. for each entry in /sproot/install/db.map there should be a file of the same name in /sproot/install/db.map. The first line of the db.map file is a checksum on the number of databases installed and listed in db.map.
– /sproot/install/vol.map should be consistent with the contents of /sproot/install/voldata. I.e. for volume each entry in /sproot/install/vol.map there should be a file of the same name in /sproot/install/voldata. The structure of the vol.map file is different from the db.map file, but the checksums in both files must be identical. Entries in the vol.map file differ for database that only consist of 1 volume and databases that consist of multiple volumes:

One volume Three volumes
DBID DBID
1 3
VOLID VOLID1
VOLID2
VOLID3

– In case the database was installed with the verification option, then there will also be a file with the extension .VFR for each volume. The first line of the db.map file is a checksum on the number of databases installed and should be equal to the checksum in db.map and the number of database ID's (DBID) and listed in vol.map.
– /sproot/install/vol.map should be consistent with the contents of /sproot/spdbvols. I.e. for each volume listed in /sproot/install/voldata and vol.map there should be a directory in /sproot/spdbvols.
– Each subdirectory of /sproot/spdbvols represent a database volume and contains pointers to the location of the actual database files. These pointer location can be listed with "ls –l" on UNIX platforms and on NT you must use the type command followed by the name of the pointer file. Please note that the pointer to a file is build up out of two parts. A general part called after the volume ID, and a second part called after the file name.
– Check if all the pointers are pointing to an existing database file.

If you found inconsistencies in any of the above you now may be able to repair the situation by deleting and editing the above mentioned files and directories. Only do this when you are positive about the source of the problem. Do not forget to make a back up first and only do this after you stopped the dbserver first.

If you feel insecure about anything mentioned above please do not use these instructions. SilverPlatter TechnicaI Support can remote login to your ERL server using telnet and make the edits for you, or alternatively you could send us a listing ("ls –l" or "dir") of the directories mentioned above including the actual vol.map and db.map file. Then we can edit the vol.map and db.map files and send them back to you with specific instructions on how to proceed.

How do I control who can access my ERL server?
Access control via IP filtering is done using the ipincl file and ipexcl file in the /sproot/cfiles directory. Ipincl is the file of included IP addresses and ipexcl is the file of excluded addresses. The IP address or subnet range is entered in the file to allow (or disallow) users with that IP address or subnet to access the server.

IP addresses are entered in the file in the following forms.

Individual address:
12.34.56.78

Entire subnet range (allows access to addresses 12.34.56.1 through 12.34.56.255):
12.34.56.0
If subnet masking is implemented at your site, the subnet mask can be included after the IP address in the file as follows:

12.34.56.0 255.255.255.0 (filtering based on first three octets of address)

12.34.0.0 255.255.0.0 (filtering based on first two octets of address)

NOTE: Entries in the ipincl file should be left–justified. Leading spaces in front of the IP addresses will cause the ERL server to ignore that entry.

How do I make changes to ipincl and ipexcl files?
The /sproot/cfiles ipincl and ipexcl files can be edited directly using a standard text editor (vi, notepad, etc.) or they can be accessed from within the WebAdmin administrative program by clicking on Databases, then Database Menu.

What syntax will not work in the ipincl or ipexcl files?
Addresses using asterisks to indicate a range of addresses (12.34.56.*) and those that use a hyphenated range of addresses (12.34.56.1–99) will not work correctly.

What happens if a user attempts to connect to the server from an unauthorized IP address?
Users whose IP addresses are not listed in the ipincl file, or whose addresses are included in the ipexcl file will see the following error after entering a username and password:

"Error from server [name] EC_DENIED"

The message may also include the text "Client Address not Authorized"

Where can I find documentation on the ipincl and ipexcl files?
The ipincl file in /sproot/cfiles contains example syntax. There is also documentation on these files included in the ERL 4.0 Administrator's Manual. If you are using WebAdmin, clicking on the yellow tab containing the question mark will access the feature–specific help for IP address checking.

The default ipincl file from sproot/cfiles directory contains the following:
# Enter at the bottom of this file the IP addresses of computers that
# should be allowed to access this server. This includes SPIRS clients,
# WebSPIRS, WebAdmin, and the Z39.50 Gateway.
#
# After editing, save the changes and close the file.
#
# Entries are of the form,
#
# IP Address IP netmask
# or
# IP Network IP netmask
#
# The IP Netmask is optional.
#
# Please note the boundaries for the various different network classes.
#
# Class A : 1 –> 127 ( 8 bits for network, 24 bits for host. ).
# Class B : 128 –> 191 ( 16 bits for network, 16 bits for host ).
# Class C : 192 –> 223 ( 24 bits for network, 8 bits for host ).
#
# Simple Class A address = A
# Simple Class B address = A.B
# Simple Class C address = A.B.C
#
# Where A is in the ranges defined above for the given class.
#
# If WebAdmin, WebSPIRS or the Z39.50 Option are to run on this computer,
# localhost (127.0.0.1) must be allowed.
#
# Examples:
#
# 123.45.67.89 255.255.254.0
#
# This would allow access to all addresses on either of the two class
# C networks; that is, all addresses from 123.45.67.1 to 123.45.67.254,
# and all addresses from 123.45.66.254.
#
# If there is no netmask next to an IP address, then that IP address is
# sufficiently specific as in the following:
#
# 123.45.67.0
#
# This would allow access from 123.45.67.1 to 123.45.67.254 (an entire
# Class C network).
#
# 123.45.67
#
# This is equivalent to 123.45.67.0.
#
#
# 0.0.0.0 0.0.0.0
#
# This gives access to the whole world.
#
#
# Add your IP addresses here:
127.0.0.1

How do I implement the functionality that was achieved using the /sproot/files/offsite file in ERL v2.x when using ERL v4.x?
The offsite file is not longer used in ERL v4.x. You may allow a user to login regardless of his IP address by creating a user specific ipincl file within the /sproot/cfiles directory structure. The ipincl file should include a lien that reads:

0.0.0.0 0.0.0.0

This entry will allow the specific user to access the ERL server from any IP address.

Where is the dbmenu.cfg file?
ERL v2.x: The dbmenu.cfg and dbmenu.eg file are located in the /sproot/files directory.

ERL v4.x: The dbmenu.cfg and dbmenu.eg are located in the /sproot/cfiles subdirectory.

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

Configurable parameters in the erlsrvr.cfg file (ERL v4.x).
This section allows you to configure the intervals for statistics files to be read or split.

/*
/* statistics config
/*

The statdump parameter places a "bookmark" in the file so that the entire file doesn't have to be read when generating a statistics report. The statdump value governs how often this is done.

/* statdump – how often internal status is stored

The acceptable values for statdump are as follows. These govern at what interval the action is performed. Hourly will perform the action hourly, daily will perform the action daily, etc.

/* never, yearly, monthly, weekly, daily, hourly, twomin. statdump = daily

Setting the following parameter will retain the temporary stats files generated by dbserver.

/* statkeeptempories
/* – set to 1 to keep temporary isam files generated by a stats request.
statkeeptempories=0

Dbserver has various options to retain debugging information. These are governed by the following section:

/*
/* debugging keys
/*
Show debugn is set to zero by default. Set the value to 1 to enable portfolio debugging. Same values are true for core debugging.

/* show_debugn = 1 for portfolio debugging
show_debugn = 0
core.debug = 0

This parameter governs the way that a core file is produced. Setting it to 1 allows the server to clean up processes before producing a core file. Setting the value to zero will produce the core file without cleaning up processes first.

/* corefile = 1 means produce a core file after cleaning up i.e. normal
/* corefile > 1 means produce a core file without cleaning up i.e. loop on exit corefile = 1

Setting a filename for the msglog parameter will record the last connection to the server. The value can be set with a wildcard (*) to log the dxp message information for all logins.

/* msglog = for DXP message logging of last connection
/* msglog = * for DXP message loggin of all connections
/* msglog = * for DXP logging of each connection in/logs

The dxp logging information can be output to a file by including a filename at the following parameter:

Setting the raw log name writes all the dxp logging information to a file.
/* rawlog = for raw DXP message log
/* rawlog = raw.log

For Windows NT, this provision allows a user to enter the debugging program when a fatal error occurs. To enable this feature, set the debug_break parameter to 1 and provide a user name next to the login_break parameter.

/* On NT: debug_break = 1 breaks into the debugger should we get an assert
/* login_break is the name of a user ID to break into the debugger debug_break = 0
/* login_break = break
This section controls what information is displayed when the listconn command is executed from the sproot/bin directory. Setting the value of 1 displays the IP address of the user connecting to the server. Setting the value to 2 displays both the IP address of the user and the IP address of the WebSPIRS gateway they are accessing through.

/* show_ipaddress = 0 for ERL 2.x style IP address reporting
/* i.e. displays client address or gateway address
/* show_ipaddress = 1 displays client address only
/* show_ipaddress = 2 displays both client and gateway addresses show_ipaddress = 0

This parameter is the value, in minutes that a user can remain connected to the ERL server without any searching activity. The default value is 30 minutes. This means that if a user performs no searches for a period of 30 minutes, they will be logged out.

inactivelimit = 30

This is the tcp/ip port that the erl server uses to communicate. The port is used for inbound and outbound communication and should be open for traffic in each direction.

tcpip_port = 416

This section allocates memory for each user connection. Thus the values set are PER CONNECTION rather than summed for all connected users.

[dbsconn]
MM.SPILL_DEVICE = NONE

The REQUIRED parameters automatically allocate this amount of memory, in bytes, for each user connection the server is configured for, whether or not any users are connected. Thus it is recommended not to raise this value.

/* minimum memory to start a connection
MM.VM_REQUIRED = 2097152
MM.FAR_REQUIRED = 2097152

The DESIRED parameter will only allocate this amount of memory, in bytes, if the memory is available on the server.

/* maximum memory a connection can use
MM.VM_DESIRED=8456000
MM.FAR_DESIRED=8456000

How do I change the ERL server information?
For UNIX Versions of ERL:

ERL contains a script that allows the administrator to change server information. Data that was entered during the ERL server install can be changed by running this script. The following steps should help you execute it:

1. Shutdown the ERL Dbserver software.

2. Login to the ERL server as user ERL.

3. Change to the /sproot/scripts directory.

4. Enter the command: ./changeservinfo

5. Follow the on–screen instructions for changing server information and saving the changes.

The changeservinfo script allows you to change the information for Server ID, Server name, Organization ID, Organization name, Site ID, Site name and Maximum logins.

For NT version of ERL

You must stop the ERL server and reinstall the same version of the software.

1. Download the appropriate version of ERL from our FTP site. It can be located at:
ftp://ftp.silverplatter.com/software/erl–server/nt/

2. Stop the ERL service from the "NT Service Manager" Under My Computer, Control Panel, select the "services" icon. Next, find the service named "ERL Server" and mouse click the "STOP" button from the menu.

3. Make a backup of your \sproot directory.

4. Click on the installation program "erlsrvr.exe" that you downloaded.

5. Follow the on–screen instructions and make any changes you would like.

This process will allow you to change information for Server ID, Server name, Organization ID, Organization name, Site ID, Site name and Maximum logins.

How do I change the order in which databases display on the SELECT DATABASES window of WebSPIRS, WinSPIRS, or MacSPIRS?
If you are using ERL you can create a menu that categorizes databases in SPIRS clients. You can also define the order of database groups within a category, and the chronological order of databases within a database group.

Using a text editor

1. Open /sproot/cfiles/dbmenu.eg in a text editor. (If you are running ERL v. 2.x this file is located in sproot/files.)

The file contains the following components.

Dbmenu Heading Description

[Sn] Sequence order [S1], [S2], [S3], etc.

MenuText= Heading; displayed in SPIRS if the user is authorized to at least one database within the menu section.

Indent= Indentation level, starting with 1. (0 is reserved for server names.)

Description= Explanatory note; shown in SPIRS if the user asks for a description.

Databases= List of two– and four–character codes of the database groups you want in the section.

Key= Text string beginning with "$"; not used by SPIRS, and not required.

[order] Determines the order of databases within a database group. In the example that follows, the line "AS=AS7A,AS7B" displays the ASFA databases in chronological order. Note: There is only one [order] section in the file. You can list all the database groups for which you have more than one database.

2. Modify the file to include the desired headings, databases, descriptions, and order.

3. Rename the file as dbmenu.cfg, and save it in the same directory.

4. Unmount and remount a database. This will cause dblbuild.exe to read the dbmenu.cfg file and update the menu.

For example, you can create a menu that appears in clients as below:
State University
Medical
MEDLINE (R) Advanced 1996–1997
MEDLINE (R) Advanced 1995–1996
MEDLINE (R) Advanced 1994–1995
Psychological
PsycLIT 1988–1992
Environmental
ASFA 1988–1996
ASFA 1978–1987
E–CD 1973–1988

Sample Menu in Editor
The sample menu above appears in a text editor as follows:

[S1]
MenuText=Medical
Indent=1
Description=Medical databases
Databases=MEDS
Key=$medical
[S2]
MenuText=Psychological
Indent=1
Description=Psychological databases
Databases=PY
Key=$psychological
[S3]
MenuText=Environmental
Indent=1
Description=Aquatic Science databases
Databases=AS,C1
Key=$environmental
[order]
MEDS=MEDS008A,MEDS008B,MEDS008C
AS=AS7A,AS7B

Please note in ERL version 4 related software bug 8285 at the following URL: http://www.ovid.com/site/support/8285.jsp

My library does not subscribe to all Publishers in SilverLinker. Can I exclude these publisher sites?
To exclude a publisher site, it is necessary to edit the erlsrvr.cfg file located on the ERL server under the sproot/bin directory.

Add a [silverlinker] heading to the erlsrvr.cfg file above the /*SilverLinker configuration line.
Beneath [silverlinker] add exclude=
followed by the publisher URL.

The hostname must include the protocol prefix, http:// and the trailing slash character (/).

Separate multiple hostnames with a space character. An example is below:

[silverlinker]
exclude=http://www.journalpub.com/ http://www.europe.anypub.com/
Please note you should list all of your exclude links on the same line, it may get very long but it will only work in this way, each URL that is entered must end with the /

This excludes links from two hypothetical sites, the European mirror site of anypub.com and the main journalpub.com site.

Many publishers provide "mirror" sites, which duplicate their online content at a location physically close to their customers. You can use the exclude facility to ignore geographic mirror sites that are not appropriate for your location.

Fatal error: 64756u Assertion failed: res == EXIT_SUCCESS, file erlif.cpp.....
This problem occurs when a customer attempts to use the new Wilson discs that SilverPlatter distributes on CD–Recordable discs on AIX systems running ERL.

In order to fix this problem a patch is required for the AIX operating system which can be downloaded from the IBM website.

How do I mount a CD (in Solaris) so I can copy database files for a magnetic link install?
Execute the following command from a prompt to mount the device and display all the files in uppercase characters:

mount –F hsfs –r –o nomaplcase [device_name] or, mount –F hsfs –r –o nomaplcase /dev/dsk/c0t6d0s0

The ERL service will not start on Windows NT
"Cannot start service, service is disabled"

Click on Control Panels/Services to bring up the services control manager. Highlight the ERL service and click on "HW profiles". The service profile window is displayed. Click on the "Enable" button to enable the service.

"Cannot start service due to a logon failure"

Click on Control Panels/Services to bring up the services control manager. Highlight the ERL service and click on "Startup". In the "Log on as" section, click on "System account". Close the dialog box and click on the "Start" button to start the service.

Uknown Error code FBA1––––In UNIX
If the above mentioned error occurs while trying to start dbserver you should run pkgrm for ERL then pkgadd.

The error may be caused by missing files essential to the dbserver startup process. In most cases these files are removed by accident, but once you run pkgrm and pkgadd for ERL this will restore all missing files.

"Usage statistics task failed to initialize."
The ERL server contains a dbserver.log file in the /sproot/bin directory. When statistical files become corrupt, this file can contain errors such as:

nassert (f–>ateof || recsum((ulong *)f–>recbuf,f–>pif–>dreclen) == 0) failed – isam.c(144)

or:

Usage statistics task failed to initialize

When these type of errors are encountered, the best option may be to back up and clear the statistics files. To back up the statistics files, change to the /sproot/files directory and enter:

cp dstats.dat dstatsdat.org
cp dstats.idx dstatsidx.org
cp pstats.dat pstatsdat.org
cp pstats.idx pstatsidx.org

To clear the ERL statistics completely, copy the statistics 'masters' files into the working copies. This can be done by changing to the /sproot directory and entering:

cp masters/dstats.* files
cp masters/pstats.* files

Connection Lost message from ERL 4.03
If SPLHA.* files reside on the sproot/cfiles directory or subdirectory of your ERL server and a connection lost message occurs it may be caused by the way the files were copied to the cfiles directory.

There are two ways to correct this problem:

1. FTP the files again in binary mode to the cfiles directory.

2. Run the hfix.pl utility found in the HCU directory to correct the file format. The syntax to invoke the hfix.pl is as follows:

/sproot/perl/bin perl hfix.pl –I /sproot/perl/lib

This license has FAILED authentication. Do you wish to retry this sheet (y/n) [n] ?
Areas to check when having difficulty installing/authenticating an ERL license sheet.

If a license is not being authenticated, the problem may lie in the "Server ID" entry. The license sheet information shows SP(When XXX is the Server ID). When entering this information within dbserver, DO NOT INCLUDE THE SP$ when entering the Server ID.

When installing an ERL license, "License Manager" will run an algorithm to authenticate the license sheet. During this process the fields below are compared to validate the license sheet.

License ID
Databases
Server ID
Maximum number of users
Allow magnetic install
Expiry date (dd/mm/yy)
Cost per record without abstract
Cost per record with abstract
Cost per Kb of characters
Cost per minute of connect time


If the above fields are not consistent, then you will receive the following error message:

This license has FAILED authentication.
Do you wish to retry this sheet (y/n) [n] ?

Be sure the enter the information exactly as it appears on the license sheet.

Fatal error: status = LastInSet: 160
If a similar message appears on the server screen when attempting to start the ERL process

ERL database server starting...
606: Started ERL Database Server v4.03
606: Fatal error: status = LastInSet: 160
/sproot/scripts/start_server[8]: 606 Abort

The problem can be resolved by removing the stats files (001.dat/idx) in /sproot/erlstat.
IMPORTANT NOTE: This means you will lose statistics for the period up to and including the day you remove these files.

nassert f–>ateof || recsum((uint4 *)f–>recbuf,f–>pif–>dreclen)==0)failed isam.c (114)
ERL v2.09.5: This error message appears when starting the ERL server and may be caused by corrupt users.dat and users.idx files located in sproot/files.

Solution: Replace these files with either a backup copy of user.dat and users.idx or recopy them from the sproot/masters directory.

IMPORTANT NOTICE:

If you replace users.dat and users.idx with copies from the sproot/masters directory you will need to recreate all ERL user accounts with ERLAdmin.

DBServer fails to start – fatal error bind 13: permission denied and spsystem: tried to execute non–root script:
./cdmount /dev/cd0.
Trying to start the ERL DBServer at the UNIX command prompt and the following error appears:

# dbserver –m
54158: Started ERL Database Server v4.06
54158: spsystem: tried to execute non–root script: ./cdmount
/dev/cd0
/sproot
# 33346: Fatal Error: bind: 13 – Permission denied

14000: Started dblbuild

The DBServer did not start. When doing command ps –ef|grep db all that is showing is the dblbuild with no main server process running.

An error will also appear when trying to mount a new database. The error will show as:

spsystem: tried to execute non–root script: ./cdmount /dev/cd0 This is not the right CD (Volume not mounted)

These error message indicates that file permissions are incorrect in /sproot/bin correct as follows:

cd /sproot/bin
chmod u+s /sproot/bin/cdmount
chmod u+s /sproot/bin/cdumount
chmod u+s /sproot/bin/dbserver
chmod u+s /sproot/bin/dbinstall

The 's' in the file permissions is very important. This is called a setuid bit. It can be set with the following command: chmod u+s cdmount

To change owner please do as follows:

cd /sproot/bin
chown root cdmount
chown root cdumount
chown root dbserver
chown root dbinstall

Once the file permissions have been corrected, the administrator should be able to start the ERL server and to mount a new database on the ERL server.

Error when installing NU9A (SP–035–103) on ERL dbserver for SCO UNIX
Verification Manager (E) : Failed to generate data for NU9A
Database Manager (E) : Failed to install database
Database manager (E) : 'CINAHL (R) Database 1982–1996'
Install failed

There is a problem with a directory name on this CD. To successfully install the database, do not 'store verification data' during the installation.

A replacement NU9A disk (SP–035–106) will be sent out to all subscribers to eliminate this problem. If you successfully install NU9A by choosing not to store verification data, you are not required to install the replacement NU9A.

I am receiving the following error in my ERL dbserver log file; FBB0 IPC_QOPEN and my Unix ERL dbserver is shut down.
SEMAPHORE, IPC, QUEUE errors have to do with shared memory in Unix. At the Unix prompt, type:

ipcs

To get a console display of the shared memory information as follows:

IPC status from as of Tue Oct 27 15:01:55 1998

T ID KEY MODE OWNER GROUP Message Queues: q 176 0x0b090307 ––rw–rw–rw– erl erl q 1 0x0b090341 ––rw–rw–rw– erl erl Shared Memory: m 1300 0x0f090307 ––rw–rw–rw– erl erl m 1101 0x10090307 ––rw–rw–rw– erl erl m 1102 0x11090307 ––rw–rw–rw– erl erl m 3 0x0f090341 ––rw–rw–rw– erl erl
Semaphore facility not in system.

Since erl is shutdown, you may get rid of (clean up) all entries owned by erl. To clear a line, you would type:

ipcrm –q 176

or

ipcrm –m 1300

Delete each entry owned by erl. Then restart ERL dbserver. This should get rid of ghost processes.

I just upgraded to AIX 4.3 and now I get errors trying to open databases with any client.
SilverPlatter's ERL DBServer (version 4.0x) is supported and tested on AIX v4.2.1. After upgrading to AIX 4.3, the problem shows as a connection lost after choosing databases to search. The connection lost message comes every other time a database tries to open. A message may also show as Unable to find the database. The error that will report on the console of the ERL system is Cleaning up after crashed connection handler.

This issue will be corrected on the next release of ERL, version 4.11 (which will be supported and tested on AIX 4.3).

A temporary fix is to comment out the SilverLinker line in the erlsrvr.cfg file. If SilverLinker is not used, this will not present any problems. If SilverLinker is used, ERL must be run on AIX 4.2.1.

To correct go into erlsrvr.cfg and comment out the following line: provider.SLINK=perlprov.

Symptoms: Clients connecting to an ERL 4.11 server receive no response, and the error "nassert (size < vmtotal || dontdie) failed" appears in the dbserver.log
Problem: This error may be caused by a corrupted dbl.cfg file. Dbl.cfg is the file used by the server to present the database selection list to the clients.

Solution: Remove or rename the dbl.cfg file from the /sproot/files directory, then dismount any database and remount it. This will cause the dblbuild process to generate a new dbl.cfg file. At that point, clients can connect and search and the error is eliminated.

Error messages when using dbserver menu:
– DbInstall Warning : did not load successfully

This is the first part of the message that describes an initialisation problem encountered by the dbinstall process that is launched when option 6 from the dbserver user interface is chosen. This warning is always followed by "Initialisation (E):"

– Initialisation (E) : Failed to initialise database manager

Caused by a corruption in /sproot/install/dbdata or voldata. Usually caused by a problem during the last database installation. For more information please refer to the FAQ for database installation.

– Initialisation (E) : Can't initialise disk manager

This occurs when there is no sub–directory under /sproot/nodes or there is a file under /sproot/nodes. This error message also applies to lmadmin, the licence manager.

– Initialisation (E) : Failed to initialise volume manager

This indicates that /sproot/install/voldata/vols.map is bad or an entry in /sproot/install/voldata is bad.

– Initialisation (E): Failed to initilalise mount manager

Problems were encountered in accessing the database files. This doesn't seem to occur in practice.

– Dbserver menu starts looping

Usually caused by an empty database file (0 bytes). Check if any files or symbolic links in /sproot/spdbvols is empty, or pointing to an empty file.

Error messages in /sproot/bin/dbserver.log:
– Cleaning up after crashed connection handler 2373, session 1G.7Z:K1

These are new messages with dbserver 4.0. In previous versions a crashed connection went unreported unless an assertion occurred. The session–id refers to the session id against which statistics are logged. As for why the connections are crashing can not be determined from this error message.

– Login user demo01 from 206.243.199.38 failed: invalid username or password

This is the default message that is written to the dbserver.log file when a user is using an incorrect username password combination when trying to login to the ERL dbserver.

– Login user demo01 from 206.243.199.38 failed: client address not authorised – Login user mg from (null) failed: Client or gate way address not authorised: 194.192.186.230

This is the default message that is written to the dbserver.log file when a user is is trying to login to the ERL dbserver from an IP–address that is not authorised by /sproot/cfiles/ipincl and ipexcl, or by the userspecific IP address checking file for this user.

– nassert (!((ma–>usemode & cARRAY_READONLY) &&Subscript >= (ha–>elementCount))) failed – array.c(1389)

This is a SEARCH ENGINE (=CORE) error message and is usually the result of a previous error.

– nassert (from != mmHNULL) failed – mmflat.c(2568)
This is a SEARCH ENGINE (=CORE) error message and indicates a problem with the internal data of the server and may be the result of a previous error. It may also indicate a lack of virtual memory for the server to accommodate a database search request. Increasing the MM.VM_DESIRED setting in the [CORE] section of /sproot/bin/erlsrvr.cfg may resolve the problem.

– nassert (hitkey–>rkey.RecordNumber == GETL(ABSREC) + dtp–>lrecno – 1) failed – hits.c(337)

A problem caused by a combination of marking and reading of data which can cause pre ERL 4.04 servers to use an incorrect record number when fetching data.

– nassert (eptr != NULL) failed – separse.c(766)

This is a SEARCH ENGINE (=CORE) error message and is usually the result of a previous error.

– dbserver.c(706)
Fatal error: Fork error, errno = 12 Fatal error: Unknown error, code FBB8

Out of memory. In general if an unknown error code is printed you may be able to discover the error message that should have been printed by doing: grep FBB8 /sproot/bin/*.err

– 2 to 3 decompression given an odd length or: 0105 cErrorOddDecompressLength nassert (!hCache[i].inUse) failed – stndbtre.c

This is a SEARCH ENGINE (=CORE) error message and indicates a bad database file. The database file on the cd was bad or a file failed to copy properly and was not noticed by dbinstall. This error is more common on NT using DVD drives.

– Started ERL Database Server v4.03
Started dblbuild
Fatal error: Unknown error, code FBB3
Fatal error: Unknown error, code FBB3
nassert (from != mmHNULL) failed – mmflat.c(2568)

– nassert (hinfo–>messageHandle) failed – holdings.c(1690)

Problem with holdings files in /sproot/cfiles, possibly corrupt.

– Started ERL Database Server v4.03a
Started dblbuild
Fatal error: Unknown error, code FBA0

User tries to start dbserver –m as root user.

– Fatal error: FC02 mmENOFARMEM – Not enough far memory available.

The server ran out of virtual memory

– ERL database server starting...
606: Started ERL Database Server v4.03
606: Fatal error: status = LastInSet: 160
/sproot/scripts/start_server[8]: 606 Abort

It means the statistical files in /sproot/erlstat are corrupt or cannot be written to due to a full filesystem.

– nassert (dbl–>ActivateCount > 0) failed – ldblist.c(217)

SEARCH ENGINE (=CORE) error message.

– Failed to load xxxxx library

The xxxxx library file indicated by component.xxxxx in erlsrvr.cfg could was not found or was incompatible with the server version, where xxxxx is one of holdings, perlprov, erlstat.

– ERLdbclose fail

Usually caused by an earlier error.

– Error writing login statistics
– Error writing logout statistics
– Failed to close statistics

Statistics failed to access the files in /sproot/erlstat maybe due to a full filesystem or file corruption.

– WARNING: can't get portfolio

The server has tried to access a portfolio which was previously referencing a database but the portfolio no longer exists. This is probably because the admin user has deleted the portfolio and has not mounted a database since that change was done and therefore the server does not know the database is no longer referenced.

– WARNING: matched portfolio ignored

The user is authorised to too many portfolios for a database and the portfolio is being ignored. The limit is set at 1024 in ERL 4.

– Abort (looping)

Only happens when certain debugging options have been selected in erlsrvr.cfg and indicates that a connection crashed in debug mode.
– Got User 1 Signal – looping

Connection received a kill –1 and has entered a looping state. Used for debugging purposes.

– Fork error:

Out of memory

– config file item xxx value xxx not recognised

Unknown configuration found for statdump in /sproot/bin/erlsrvr.cfg

– Connection handler xxx failed to cleanup

Server waited for connection to stop before shutting down but connection did not stop.

– Connection handler xxx would not die

Server forcibly killed conenction handler that would not stop when shutting down.

– xxx task (yyy) failed to cleanup
– xxx task (yyy) would not die

Server waited for dblbuild or user interface to stop before shutting down but they did not stop.

– Error: bad coverage date: xxx
When showing the mounted databases the user interface found an invalid coverage date.

Error messages in \sproot\bin\dbserver.log On NT only
– Cannot access drive \\gandalf\h\, error 53
Suggest you remove this from file d:\sproot\files\devices

This is a warning and simply means that the device is in \sproot\files\devices but wasn't found. It can be ignored, but to prevent it from coming up you should edit \sproot\files\devices.

– Started ERL Database Server v4.03
Process 332 not found in IPC_Awaken

Usually means that dblbuild didn't start because the server was not shutdown properly.

– Started ERL Database Server v4.03
Cannot set event in process 109 in IPC_Awaken, error 6

Usually means that dblbuild didn't start because the server was not shutdown properly.

– Fatal error: Error 170: OpenDesktop failed in Qthread

Usually occurs when a telnet application is trying to run the user interface and does not have the necessary permission to access the server environment. ERL 4.0 seems to have broken the ability to telnet to NT and run the user interface anyway.

– nassert (cptr != NULL) failed – .\nwapi\socket\sharedso.c(302)

Probably caused by an earlier error.

– IPC_WriteQ to CONNQ (2) – queue not found

Probably caused by an earlier error or because one of the dbserver processes has died.

– Cannot create cleanup thread.

internal error.

– Error waiting for connection handler event, error xxx

server failed to tidy up a crashed connection

NASSERT error messages for dbserver 2.09.5 and 4.0*:
The following are assertions which may occur within the server code. It does not include assertions which originate within the dbserver search engine (CORE).

scsts.c
nassert(reclen + 4 + Rs.msglen <='' csts_maxresp) data to return to client is too large to fit inside message buffer. internal error.
nassert(ip–>len == 5); Data to return to client has a data item whose length does not match data item type. Internal error.

dbl.c
assert(p < scratchpad + spadsize);
String too long too return for 'database list information', must be less than 2048 bytes.

erl.c
assert(dbh > 0 && dbh < ndbs && erldbtable[dbh].mslot ='=' 0);
Attempt to open more than 100 databases or a database already open.
assert(pdbi = getdbinf( Dbinff, dbid));
Error reading database information for the database in dbinfo.dat during database open.
assert(edbp–>ush != –1);
Error opening totals.dat file during database open.
assert(gtm–>dbh > 0 && gtm–>dbh < ndbs && erldbtable[gtm–>dbh].mslot != 0);
Attempt to open more than 100 databases or a database already open.

gdbi.c
assert(p < scratchpad + spadsize);
String too long to return for 'get database information', must be less than 4096 bytes.

hits.c
assert(hitkey–>rkey.RecordNumber == GETL(ABSREC) + dtp–>lrecno – 1);
Discrepancy with the absolute record number during 'set hits'. Possible client error.

login.c
nassert(strlen(salt) < dxs_salt_len);
Internal error during logging in using DXP challenge protocol.

portfolio.c
assert(empty != –1);
Attempt to login to a portfolio when there are no free portfolio login slots. One slot is required per portfolio per user. e.g. 10
users each using 3 PORTFOLIOS require 30 slots. Limit is 256 (MAXPFOLIOS)? in use.
assert(use);
Attempt to logout of a portfolio which is not in use. assert(count >= 1);
Internal error during accessing date string in portfolio or database.

msgloop.c
assert(ep = MALLOC(sizeof *ep));
Unable to allocate memory.
assert(ep = REALLOC(ep,size));
Unable to allocate memory.
assert(ep = MALLOC(sizeof *ep + sizeof *mp));
Unable to allocate memory.
assert(GetBEWord(buf + CSTS_MAGIC) == 0xabcd);
Error in DXP message header from client.

readdata.c
assert(dtp–>cursor.posn_method != POSN_RE);
Conflict in positioning to database record. Possible client error.

rollup.c
assert(psp–>Enddate <='' ed);
Conflict between today's date and record in pstats.dat file.

usif.c
assert(msg.len <='' maxqmsgsize);
Internal error. Message to send to usage statistics process is too large.
assert(msg.len == nbytes + sizeof type);
Internal error. Message received from usage statistics process is too large.
assert(msg.fromq == UstatQID);
Internal error. Message received from usage statistics process has unexpected message
queue identifier.
assert(msg.type == type);
Internal error. Message received from usage statistics process has unexpected type.

ustat.c
assert(msg–>len == sizeof *sp);
Internal error. Discrepancy in usage statistics process 'update statistics' message size.
assert(msg–>len == sizeof *osm);
Internal error. Discrepancy in usage statistics process 'open' message size.
assert(msg–>len == sizeof *gsm);
Internal error. Discrepancy in usage statistics process 'get statistics' message size.
assert(msg–>len == sizeof dbh);
Internal error. Discrepancy in usage statistics process 'close statistics' message size.
assert(msg–>len == sizeof *utm);
Internal error. Discrepancy in usage statistics process 'update record totals' message size.
assert(dbh > 0 && dbh < elements(stable) && stable[dbh].refcnt);
Internal error. Invalid database id in usage statistics process 'get record totals' message.
assert(msg–>len == sizeof sid + sizeof did);
Internal error. Discrepancy in usage statistics process 'update statistics' message size.

uuinfo.c
assert(isamget(Userf,1));
Error reading user information in users.dat.

dbserver.c
assert (db_file != NULL);
Error opening dbinfo.dat file.
assert(cFALSE);
Internal error starting dblbuild, user interface, usage statistics, or connection handler.
assert(countdown == 0);
Internal error shutting down server.

uif.c
assert (db_file != NULL);
Error opening dbinfo.dat file.
assert (n_mounted_dbs < sizeof (mounted_db) / sizeof(mounted_db) – 1);
Too many mounted databases (>100).
assert(isamget(db_file,2));
Error reading information from dbinfo.dat file for database being updated to.
assert(flushmnt(db_file,mntp + mslot,0));
Error writing information to dbinfo.dat file for database being updated from.
assert(msg.len <='' maxqmsgsize);
Internal error. Message to broadcast to users is too long.
assert (!fs_initialised);
Internal error. Discrepancy in status of file system intialisation. assert(b);
Error reading information from dbinfo.dat file for database being mounted.
assert(flushmnt(dbinff,mntp,0));
Error writing information to dbinfo.dat file for database being unmounted.

dblbuild.c
assert(count >= 1);
Internal error during accessing date string in database.

cdstuff.c
assert (p [0] == '/');
sproot directory in not under /.

ipcapi.c
assert (ObjHdrPtr–>Flags & HASSEMAPHORE);
Internal error. Discrepancy in shared memory object while attempting to lock or unlock it.
assert (ShmSegPtr–>Signature == SHMSIGNATURE);
Internal error. Discrepancy in shared memory segment during segment creation.
assert (i == 0);
Internal error. Failure to set signal mask or signal action during initialisation.

cstslog.c
nassert (reclen >= 4);
nassert (reclen >= 2);
nassert (reclen >= 1);
nassert (reclen >= 1);
Internal error. Discrepancy in record length when writing message response record to message log file.

isam.c
assert(pix = VALIX(f,ix));
Invalid key used to read record from ISAM file.
assert(f–>ateof || recsum((uint4 *)f–>recbuf,f–>pif–>dreclen) == 0);
End of file found or record checksum failure in reading from ISAM file.
assert(nsegs <='' pix–>inumseg);
Internal error. Discrepancy with search key and keys in ISAM file.

Repetitive failing attempts to login as admin user may be caused by Z–gateway!
When you notice a repetitive error like "Login user admin from 127.0.0.1 failed: invalid username or password" on the console or in the dbserver.log this may be caused by a Z–gateway. Once it starts occuring, the message continues to log every 10 seconds until the process number associated with this error is killed off.

When you are running the Z39.50 gateway you will find the following two lines in the /sproot/z3950/erlsrvr.cfg file.
ERL.USERNAME = admin
ERL.PASSWORD = admin
These are the username and password you specified during the installation to be used when building the Explain database.
The default username is "admin".If you change the password for "admin" in WebAdmin and you don't change the password specified in this file, you will keep on receiving the "Login user admin from "IP of Z gateway" failed:invalid username or password." message.

Error message when trying to start ERL server
$ ./dbserver –m
327: Started ERL Database Server v4.04
327: Fatal error: isamopen(/sproot/files/dbinfo.dat) failed, isam_err:
40, sysio cod: 0 Abort $

This means that the dbinfo.dat/.idx pair of "ctree" files have become corrupted. This may happen for instance after an unexpected power down of the server. It can also serve as an first indicator that the hard disk drive on which /sproot is located is about to die. You are strongly advised to make regular backups of all the .dat/.idx file pairs located in /sproot/files. These files keep most of the ERL administrative detail about; users (users.dat/idx), licenses (pfolio.dat/idx), databases (dbinfo.dat/idx), authoristations (authlist.dat/idx) and active connections (dbconn.dat/idx). Note: when you make back–ups of the .dat/idx files, you are advised to stop the dbserver first. If any of the previously mentioned ctree files have become corrupted you can restore the most recent backup to recover.

Solution in case dbinfo.dat/idx have become corrupted and you do not have a back–up:
1.Remove /sproot/files/dbinfo.dat and .idx or rename/move them to another folder.
2.Copy new dbinfo.dat and .idx files from the /sproot/masters directory to the /sproot/files directory. These new files contain no information.
3.Re–start the ERL server
4.When invoking the dbserver menu you will notice that all DB mounting/enabling information got lost. The information about installed databases (sub menu "list installed databases") should be still there. If not, more than only dbinfo files have been affected.
5.You can now mount all databases again, by using the asterisk wildcard. Do not forget also to enable the databases, and after this they should appear in the database list of the clients. 6.As a final check try to access and search one or more databases.

Solution in case dbconn.dat/idx have become corrupted and ERL server does not start:
1. Remove /sproot/files/dbconn.dat and .idx
2. Re–start the ERL server

If you're users.dat/idx, pfolio.dat/idx or authlist.dat/idx have become corrupted, you're only solution is to restore you're most recent back–up and redo all the edits made to users, licenses and authorisations after the time of back–up. At any time you can verify the integrity of the files by running the ctexport command on the .dat file, eg. /sproot/bin/ctexport –v users.dat. This will give a complete list of ERL users and should complete out

Why do I get the following error while installing ERL 4.0 on Solaris v. 2.6? "postinstall script did not complete successfully sproot/scripts/solcfgkrnl:test:argument expected"
The solcfgkrnl script being executed by the postinstall script is looking for the following lines in the /etc/system file:

set maxusers=128
set msgsys:msginfo_msgmni=200
set msgsys:msginfo_msgtql=200

Edit the file and add these lines at the end. The script will complete at that point, and the Solaris system configuration will not be affected since it doesn't use these lines.

How do I upgrade from ERL Dbserver 1.2 to 4.0?
This upgrade is not supported. You will need to remove the 1.2 package with the pkgrm command, remove the whole /sproot file system and start with a fresh installation of ERL Dbserver 4.0 as described in the ERL Administrator's manual. Contact your distributor or SilverPlatter Information first to obtain the required ERL Server ID and database license sheets.

What do I need to know before I upgrade to ERL Dbserver 4.0?
– There is no upgrade possible from 1.2
– Installed databases are not affected
– User database is automatically upgraded
– The password for the "admin" user will be reset to "admin" in the upgrade process.
– License portfolios are automatically upgraded
– Authorizations are automatically upgraded.
– The ERL Dbserver v4.0 software has been designed to replace ERL Dbserver 2.x
– Before you upgrade, you should check for and remove any commas from the Description and/or Reference fields of the user accounts.
– Due to enhancements in statistics, statistics from an ERL v2.x server will not be preserved.
– Before installing ERL v4.0 we recommend that you perform a system backup and export any statistics you wish to save. Check the release notes for more info.

The ERL Dbserver software does not need more disk space than 50Mb, however, you are advised to have up to 2 Gb of free disk space available on the disk where you install the ERL Dbserver. This space is required in order to store the Dbserver usage statistics and other log files that grow over time.

How do I upgrade from ERL Dbserver 2.x to 4.0?
Upgrading from Dbserver 2.09x on SCO and Solaris:

1. download the erlsrvr.pkg.Z for SCO or Solaris
2. type: "uncompress erlsrvr.pkg.Z" (if "uncompress" is not available on your system try "compress –d")
3. export the statistics you need
4. stop Dbserver 2.09.x (type "SHUTDOWN" in the main Dbserver menu)
5. remove 2.09.x with the "pkgrm" command.*
6. install 4.0 with the "pkgadd –d" command as described in the ERL Administrator's manual
7. Always start Dbserver 4.0 from /sproot/bin with the following command: ../scripts/start_server
8. When you reboot the server the Dbserver will be started automatically.

*NB: DO NOT REMOVE ANY FILES MANUALLY

Upgrading from Dbserver 2.11 on NT:

1. download the erlsrvr.exe file for NT
2. export the statistics you need
3. stop ERL server 2.11 NT service
4. remove 2.11 with the uninstall script from the ERL Dbserver 2.11 program group*
5. Remove the dbl.txt file from \sproot\bin if present
6. install Dbserver 4.0 by running erlsrvr.exe
7. "ERL Server v4.0" service can be started from the services control panel.
8. When you reboot the server the ERL Dbserver service will be started automatically.

*NB: DO NOT USE THE "ADD/REMOVE SOFTWARE" OPTION FROM CONTROL PANEL, DO NOT REMOVE ANY FILES MANUALLY OTHER THEN MENTIONED IN 5.

Upgrading from Dbserver 2.1x on AIX

1. download the erlsrvr.tar.Z for AIX
2. type: "uncompress erlsrvr.tar.Z"
3. Make sure you are logged in as "root". You should see a hash mark (#) prompt.
4. Change to the / (root) directory.
5. Stop the ERL Dbserver (type "SHUTDOWN" in the main Dbserver menu).
6. Type the following commands at the prompt (as the root user):
* cd / ; slibclean
* tar xpf full_pathname/erlsrvr.tar where full_pathname is the complete path to the location of the ERL Dbserver server package.
* cd /sproot
* ./postinstall
Always start Dbserver 4.0 from /sproot/bin with the following command: ../scripts/start_server

Nassert (vmd) failed – ipcapi.c (1935) error message in dbserver.log ERL 4.x
The error message indicates that the server cannot find or cannot read the MM.VM_DESIRED and/or the MM.FAR_DESIRED lines in the erlsrvr.cfg file located in sproot/bin. Check the following lines in erlsrvr.cfg in the [dbsconn] section and add values to these settings if they are blank :

/* maximum memory a connection can use
MM.VM_DESIRED = 4194304
MM.FAR_DESIRED = 4194304

You may add the values shown above or increase them.

Test by connecting to the ERL server with a client.

When trying to upgrade my NT ERL Dbserver software, I am receiving an error.
There is an old installation of this product. Please uninstall and try again.

I have already done an uninstall. How do I correct this problem?

Installation of ERL Dbserver is detecting information regarding the previous version from entries in the Registry. There are two entries that should be removed, through REGEDIT, in order to clear this error:

HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Services\ERLsrvr
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\ERLsrvr

Only for ERL NT: "There is an old installation of this product, Please uninstall and try again."
This only happens in case ERL has been installed before on that server and the registry entries for dbserver have not been cleared properly when the ERL dbserver software was removed. This will also happenin case you use the ADD/REMOVE software option from the Control Panel. If you remove ERL dbserver using the "uninstall" script from the SilverPlatter ERL program group this problem will not occur.

Solution:

1. Exit the set program for ERL dbserver

2. Open a Dos box or CommandPrompt window and run "regedit"

3. This will start the regedit program on your machine. (Please note you should be extremely carefull using this application) You must remove the following entry:

HKEY_LOCAL_MACHINE\Software\SilverPlatter Information\ERL Server

4. You must also remove: HKEY_LOCAL_MACHINE\system\currentcontrolset\services\ERLsrvr

5. You can now install ERL dbserver again.

How to disable the automatic start–up of WebSPIRS on Solaris
WebSPIRS 4 on Solaris and SCO is not installed as a package. Therefore, after deleting WebSPIRS from your system, the server will still attempt to start WebSPIRS at boot time. This can not do any harm but will produce an error message on the console.

To prevent this from happening you'll have to remove the start–up script called "SM0webspirs" from /etc/rc2.d.

Consecutive WebSPIRS installations or up–grades will NOT remove the SM0webspirs file, when one decides to have NO automatic webspirs startup.

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation.
http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:
Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful: 1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?

The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:


Library owns 1976 – present A HREF="http://www.ovid.com">Click here .


Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

WEBSPIRS

Overview

General

Operating System Configuration

Configuration Options

Error Messages

Installation


How does WebSPIRS v4.x communicate with ERL?
WebSPIRS consists of two processes, webspirsd and webspirs_user, which move information back and forth between the user's web browser and the ERL database server.

Here's a brief overview of how WebSPIRS communicates between the user's web browser and the databases they are searching:

1. The user submits a search request or clicks a button on the WebSPIRS toolbar.

2. The request goes to webspirsd:

If the request requires a connection to a database (such as search request), it is sent to the webspirs_user process.

If the request is for a file (for example, a WebSPIRS button graphic), it is processed by the webspirsd process.

3. The webspirs_user process takes the request from webspirsd and communicates with the ERL server. The ERL server, in turn, communicates with the appropriate databases to retrieve records, perform a search, etc.

4. The webspirs_user process then passes the response back to webspirsd that passes it back to the browser where it is displayed.

How can I view TIFF files in WebSpirs from my PC
A TIFF viewer is needed such as AlternaTIFF available free from http://www.alternatiff.com. Copy the viewer into the WINNT/addins folder on your PC and re–boot the machine.

UNABLE TO VIEW RECORDS USING "NEXT 10" BUTTON
Clicking on the "Next 10" button does not advance the record display past the first 10 records.

This error may be caused by the browser program not refreshing the cached information for the page. To eliminate the problem you will need to change your browser settings:

Netscape:
Select Edit/Preferences/Advanced/Cache and change the setting for "Document in cache is compared to document on network" from "Never" to "Once per Session" or "Every Time".

Internet Explorer:
Select Tools/Internet Options and click on the "Settings" button in the center of the dialog box. Under "Check for newer versions of stored pages", select "Automatically".

Once the settings are changed, click on "OK" to save the changes and you should be able to view records within Webspirs.

Customers have reported problems when using Netscape 6 in conjunction with WebSPIRS.
Netscape 6 has not been tested with WebSPIRS and is not a supported browser, at this time. Customers have reported errors and have experienced slowness when trying to navigate through screens and conduct searches using WebSPIRS.

SilverPlatter Technical Support suggests that customers reinstall a previous version of Netscape (version 4.76 or earlier) in order to take advantage of increased speed and error free searching.

Can I install WebSPIRS 4 on NT 4 workstation?
Yes this is possible but be aware that the number of TCP/IP connections on NT workstation 4.0 is limited to 10 (by NT)
For the latest information about requirements for WebSPIRS 4.1 see the SilverPlatter web site at:
http://www.ovid.com/site/support/erlgatewayreq.jsp

Why did SilverPlatter choose to utilize JavaScript and Frames in WebSPIRS v4.x?
We believe that the combination of JavaScript and frames offer the most appropriate available technology for achieving the functionality; navigability; and user interface improvements which our customers indicated they required in a Web–based client.

Does WebSPIRS 4.x use Java?
WebSPIRS relies primarily on JavaScript, which is in no way related to the Java programming language. However, WebSPIRS does use an actual Java applet to display the Table of Contents in a small number of SilverPlatter databases that support that feature (e.g. the EIU familiy of databases; DIF; OSH )

Do I need to have Web server software at my institution to run WebSPIRS?
WebSPIRS 4.x incorporates its own Web server, so no external Web server software is necessary. This in no way precludes the continued use of installed Web server software for accessing other web–based applications.

WebSPIRS v3.1 requires that commercial Web server software is installed on the machine that will run WebSPIRS.

Why did SilverPlatter make the decision to use proprietary Web server software for WebSPIRS 4.x?
Since WebSPIRS 4.x uses only a small part of the code available in a commercial Web server, we decided to incorporate WebSPIRS–specific Web server capabilities within the software in order to optimize performance.

Why does WebSPIRS v4.x open so many browser windows?
WebSPIRS 4.0 was designed with a frames interface that allows it to keep controls in view while the data on the page scrolls. In order to present the best interface even for those using small monitors and resolutions of 640x480, we need to display this interface in a window that has the toolbars and entry fields turned off. In order to do this, we need to open a new browser window for the main WebSPIRS interface. JavaScript restrictions keep us from closing the original browser window afterwards.

As we were designing WebSPIRS 4.0, we found that it was more appropriate to present some functions and information in a separate browser window. For example, if the user clicks on the Record number link to see a complete record, we present it in a new browser window. In that way, the user can choose to look at more than one record at once, or to continue to work with records in the main window without "losing" a record of interest. In addition, we can provide users with a quick way to see a full record, yet when they close the record and return to the results display, that display is scrolled to the exact place where they left it. This would be difficult or impossible to do if all functions took place in a single window.

We have taken several steps to reduce the problem that this design decision may cause. The original browser window displays a "Session in Progress" page. If the user cannot find the browser window running the original WebSPIRS interface, the "Bring it Back" button on the "Session in Progress" page will restore that window. When the user clicks Logout, WebSPIRS closes all the browser windows opened during the session.

Early adopters of WebSPIRS 4.x voiced concern that their users would be confused by the several browser windows; however, this concern has not been born out, as actual users have found the interface to be clear and easy to use. We will continue to monitor feedback from users as they gain more experience with WebSPIRS 4.x, and will continue evaluating the user interface for possible improvements, including the elimination of any Browser windows which are deemed unnecessary.

Why is the Browser's Toolbar (e.g. the Back button) unavailable in WebSPIRS v4.x?
Since WebSPIRS stores many variables both in the browser and in the server, it is very important that the user not use the Back button on the browser. We also found that users of earlier versions of WebSPIRS became confused when trying to choose between the Print function of WebSPIRS and the Print button on the toolbar of most browsers. Removing the browser's toolbar and location field reduced users' confusion and removed the likelihood of problems due to inappropriate use of the Back button.

BROWSER Back button with previous versions of WebSPIRS, use of the browser's BACK button resulted in serious errors within the WebSPIRS application, as the sequence of functions which it expected was compromised. For this reason, we made the decision to hide the Browser toolbar altogether in WebSPIRS 4+. Many WebSPIRS pages now have their own "Back" function. As we identify other areas within the application where a Back function is desirable (e.g. within the thesaurus function), we will consider adding these in a future release of WebSPIRS.

As for going "forward" , WebSPIRS screens always make it very clear to the user how to go forward to the next step, with the button indicating the most frequent and logical next step being highlighted.

WebSPIRS–Compatible Web Browsers
WebSPIRS 4.x has been tested using browsers running on the following platforms:

Netscape Navigator 3.01 (32–bit) or higher for Windows 95, Windows NT (4.0 or higher)

Netscape Communicator v4.x or higher for the Macintosh

Microsoft Internet Explorer 4.0 (32–bit) or higher for Windows 95 or Windows NT (4.0 or higher) only. (IMPORTANT: Internet Explorer for the Macintosh is not a supported web browser.)

In order to insure optimal performance of the WebSPIRS 4.x gateway, we recommend that it be used with browsers running on these platforms.

NOTE: The browser's JavaScript capabilities must be enabled at all times. Java must be enabled only if you subscribe to SilverPlatter full text databases that use the Table of Contents feature.

URL Referral for WebSPIRS v. 3.1
Instead of filtering on IP addresses, it is possible to filter on a secure URL. This is a web page where patrons must log in with a user name and password or come from a certain IP address in order to gain access to the page.

Using the example below, WebSPIRS will look for all users to come from the exact web page specified in the webspirs.cgi.

#!/bin/sh

PROGDIR=/opt/webspirs/bin
PROG=cgibaby

if [ "" = "http://www.secure–page.edu/secure.html" ]
# also can use: –o for mulitple url's...
then
cd
/default.sp.nextform=top.htm sp.username=xxx sp.password=xxxxx

else

cd
/default.sp.nextform=top.htm
fi

Why is my SDI for Current Contents returning duplicate records?
WebSPIRS 4.x

If both the weekly and archival editions of Current Contents are selected when the SDI is created then the SDI will return duplicate records. When creating SDIs for Current Contents select either the weekly databases – databases with coverage dates of only a week, or the archival databases, but do not select both the weekly and archival databases. In either case the SDI will return new, unique records.

The archival edition of Current Contents will contain the same records that appear in the weekly edition. However the weekly edition has a database family identity different than that of the archival edition and is considered a different database by the SDI function. SDI distinguishes between records within a database family identity, but does not recognize duplicate records from different database families.

What is the difference between an "SDI" and an "Alert"?
An Alert is a special kind of SDI.

When the SDI (Selective Dissemination of Information) in enabled. Any user can log in, will see a new set of buttons in the search history area, and can create an SDI. If a user would like to have updates on a specific search sent to his e–mail address periodically, he would first enter the information about the search for which he would like to create the SDI. Until there is information in the lower "Search History" area of the WebSPIRS screen, the buttons regarding SDIs are greyed out. Once a search is created, it may be set up as an SDI. Updated information will be sent to the e–mail address he enters when he creates an SDI as often as the administrator runs the script for SDIs.

When the Alert function is enabled, it is configured for one login name. The suggestion would be to enable alerts for a login name other than the one shared by most users. An example may be that if a majority of users login as c2222, the Alerts administrator may be given the login of c2222a. Everyone who logs in, either c2222 or c2222a will see the Alerts button on the right hand frame. However, the majority of users (that log in with c2222) cannot create an alert, they can only subscribe. It is only the person who logs in as c2222a who, when he/she creates an SDI, it appears in the Alerts area. He/she is the only one to be able to edit or remove an entry in the Alerts area. An instance when this would be very helpful would be if an instructor would like to make a search available to students, but does not want any students to be able to remove or edit the search. The students would subscribe to have the updates sent to their e–mail address, but the search itself has been preset by the Alerts administrator.

Why can't I create SDI's for some database families
Several database families are not available for creating SDI's because they do not contain an unique AN (accession number) or equivalent field in the database. The SDI mechanism stores the accession number (or equivalent value) for all the database records that already have been found at the initial creation or received in a SDI email by the user. Apart from the accession number the SDI mechanism also stores a checksum on the CITN (citation) fields from each record that was found.

Additionaly in the [ALERT] section of the webspirs_user.cfg file a number of databases is specified that do not have an unique indentifier in all database records. The default value for these exluded database is:
exclude.setid.p=AS,TS,IF,GB,TE,GG,LO,AT,AD,MN,PT,PG

"Dotted background" makes it difficult to read the text in WebSPIRS 4+
This is a 16 colour display problem caused by the graphics device of the end–user machine. If the end user machine only supports standard VGA 16 colour display, there is nothing you can about this. However on most machines you can simply adjust this setting by changing the display settings in the Control Panel.

For NT servers you will find that the 16 color display is the default install value which can be easily changed as described above.

Increasing the "Color Palette" to 256 colors or better will change the background to a more "eye friendly" style.

If my OPAC allows searching on ISSN/ISBN, how can I tell if "link to holdings" with WebSPIRS 4+ is possible?
When you are trying to implement the "link to holdings" in WebSPIRS 4+ you may find the following information useful to determine if a link is. Go to the web page from which it is possible to search the OPAC on ISSN/ISBN numbers. If you look in the source of this web page, you will notice that the actual "URL" that will bring you to the correct holding page of the web OPAC is either a URL using

Examples:
1. If you find in the source of this page: you will not be able construct the link URL from this page. You may need to talk to the people that wrote the scripts behind these web OPAC pages to ask if they can also allow searching directly from a URL on ISSN/ISBN. All they may need to do is write a new CGI script that allows searching directly from a URL on ISSN/ISBN numbers.

2.If, in the source of the page from where you start the search, you find:,you will probably be able to construct a link URL that you can use to create the "link to holdings" in WebSPIRS.

For further information on how to create the configuration section for "link to holdings in the webspirs_user.cfg file, please refer to the WebSPIRS Administrators manual.

Creating SDI's with one user license databases on UNIX......
The following are instructions on how to fix the problem when creating SDI's on databases with 1 user licenses.

Step 1
Create a file and name it "message" in the /webspirs/bin directory. This file contains the message the user will receive when they attempt to create the SDI. This is a sample message:

Subject: SDI Warning!

This email is a warning that WebSPIRS currently is unable to create your SDI.
At this moment there are no free database logins available to create the SDI baseline search for your search. Therefore the status of this SDI will remain "PROCESSING" until the baseline can be completed. Should the status of your SDI remain "PROCESSING" for more then 2 days, please contact the library help desk.

Step 2
Modify the spmail file in the webspirs/bin directory to look as follows:

#!/bin/sh
PATH=:/usr/sbin:/usr/lib:/usr/bin:/usr/local/bin
echo sendmail $2 <$1>
echo "display value 1"
cat $1
echo "display value 2"
echo $2
echo $3
var=`grep –c "" $1`
if [ $var –ne 0 ]
then sendmail $2 < message
else
if [ $# –eq 3 ]
then sendmail –f $3 $2 <$1>
else sendmail $2 <$1>
fi
fi

Step 3
Create the runbaseline.sh executable script in the cron directory. The following is the syntax for this script:

#!/bin/sh
clear
echo "*********************************************"

#Search for *.def files below the directory of execution
cd [path to webspirs/bin directory] For example:
cd /9gig/webspirs42/bin
for i in `find . –name "*.def"`

#Assign the number of occurrences with the selected database name
#in the SP.NEXTFORM line of the *.def file to the variable "var"
#If var is different from zero it will execute the sdireg.sh command"
#on this file only

do
var=`grep –c SP.NEXTFORM=baseline.htm $i`
if [ $var –ne 0 ]
then
echo "creating SDI baseline $i"
./sdireg.sh $i
fi
done
exit

The WebSPIRS v4.x Server Stops Responding with Multiple Simultaneous Users
1. Check the error.log file in the webspirs/bin directory.

2. If a message appears indicating that there are too many open files, raise the number of open file descriptors for the user account running WebSPIRS. For example, on Solaris you might type:

ulimit –n 1024

3. Save and close the file.

WebSPIRS 3.1 on Solaris is generating the following error:
User Number: 481945 Remote Address: 134.7.74.106 Remote Host:
134.7.74.106ld.so.
1: internal: malloc failed

cgichild giving up on cgiadultld.so.1: internal: malloc failed

This problem can be rectified by editing the WebSPIRS startup script and increasing the ulimit parameter, which governs how many files can be simultaneously opened by any given process.

The script that starts WebSPIRS is named "SM0webspirs" (the third character is a zero) and is located in the /etc/rc2.d directory. The following syntax can be added to the script at the beginning:

ulimit –S –n 1024

Save the edited script and restart WebSPIRS. The error message should be eliminated at this point.

WinSPIRS has a "Show Held Titles" option. Does WebSPIRS have this feature?
WebSPIRS does not have the "show held titles" option.

Something similar is to sort by LHM field. The way that you do this is:

Click on the "Change Display" option from the search screen
Click on the Radio Button in front of "Sort By"
In the drop–down box, choose LHM
Choose "Ascending"
Click on the drop down box "If fewer than ___ records" and choose 500
Confirm changes

The search results will display the records with the LHM field at the top. There is a limit on 500 records per search, due to the burden it places on the server. (If your search contains more than 500 records, you may choose to add a limit by year or otherwise limit your search.)

WebSPIRS.net not filtering IPMASKS correctly...
When editing the webspirs.net file to autologin certain IP ranges you must follow the recommended procedures in the file and confirm the following:

In the IPMASKS section include only one space in between the IP range and the section it should use. WebSPIRS will not accept a tab, or any other delimiters.

Correct Entry:

[DEFAULTCONFIGURATION]
sp.username.suggest=test
sp.password.suggest=test

[ALPHACONFIGURATION]
sp.username=admin
sp.password=admin

[IPMASKS]
000.000.000.00(SPACE)ALPHACONFIGURATION

How do I configure WebSPIRS 4.3 to logout to a specific URL?
WebSPIRS 4.3 can be configured to take the user to a specified web page upon logout. This configuration is available in the /webspirs/bin/webspirs_user.cfg file. Uncomment the following entry by removing the /* and changing the http:// information to reflect the URL you choose:

[SP.FORM]
/* last.p=http://www.silverplatter.com

How can I change the default in WebSPIRS so marked records will clear after printing?
It is easy to change this default by editing the print.htm file in /webspirs/template as follows:
Change the value="0" to value="1" in the line [SP_MACRO] sp.assign macro="sp.marked.clear.p" value="1" reset="y"[/SP_MACRO] and the 'Clear marked records after printing' box is now presented with a check mark and marked records are cleared after printing.

How do I configure WebSPIRS for a Referring URL?
WebSPIRS 4.3 has the ability to use a referring page to authenticate users. This function works in the following manner:

IF YOU ARE USING SILVERPLATTER INTERNET SERVERS:

1) Write Technical Support and provide:

a) URL you use to access WebSPIRS
b) URL at which the WebSPIRS link is located

IF YOU ARE USING WEBSPIRS LOCALLY:

1) You must create a USER subdirectory within the /webspirs/bin directory (follow instructions in the Administrator's Guide Click here for Administrator's Guide )
2) Within the USER subdirectory, create a file named "startup.cfg". The only line in the file is the following:

Sp.from=http://www.silverplatter.com/support

Where www.silverplatter.com/support is the server name or ip address and path
The Administrator of the web site needs to configure the link on the site to go to the WebSPIRS URL with the subdirectory, which you created in step 1.

Note:
If you are using this referring URL, you cannot use ip filtering. The startups.cfg file takes precedence over the webspirs.net file, where the ip filtering takes place. If you would like both ip filtering and referral, you would need two distinct URLs to access WebSPIRS.

Changing WebSPIRS SDI default limit (records.p=1000)
SilverPlatter does not recommend changing the default limit for SDIs. Increasing the default limit creates a VERY LARGE set of files on the server. If your SDI is named "test", the files that are created are test.def and test.key, located in the /webspirs/bin/ directory. There are entries within the .key file for each record in the original retrieval. If you have an original search of more than 1000 records, the .key file contains identifiers for each record (very large). Also, when the script is run against the files, in order to e–mail updates to users, the program looks at each and every entry in the .key file. This takes a long time, with many users creating large SDIs.

As an alternative to increasing limits, the following suggests a different approach. If the user appends one of the following to the prepared search, the results would be smaller in number:

XXXX and (UD>200100)

or

XXXX and (PY>2000)

(where XXXX represents their original search)

Since WebSPIRS requires that a user open all years of a specific database, in order to create an SDI, the above suggestion provides a helpful method for users who wish to create an SDI against a broad search.

How do I increase the SORT limits in WebSPIRS version 4 or higher?
The sort limits can be changed in the template chglist.htm, located in the /webspirs/template directory. You are cautioned not to increase the limits beyond 500. Changing these values affects performance.

Replace the following entries:

selections=sp.record.sortlimit.p value="50"
iftrue="SELECTED"[/SP_MACRO]>50
selections=sp.record.sortlimit.p value="100"
iftrue="SELECTED"[/SP_MACRO]>100
selections=sp.record.sortlimit.p value="150"
iftrue="SELECTED"[/SP_MACRO]>150
selections=sp.record.sortlimit.p value="200"
iftrue="SELECTED"[/SP_MACRO]>200

With the following:

selections=sp.record.sortlimit.p value="100"
iftrue="SELECTED"[/SP_MACRO]>100
selections=sp.record.sortlimit.p value="200"
iftrue="SELECTED"[/SP_MACRO]>200
selections=sp.record.sortlimit.p value="300"
iftrue="SELECTED"[/SP_MACRO]>300
selections=sp.record.sortlimit.p value="400"
iftrue="SELECTED"[/SP_MACRO]>400
selections=sp.record.sortlimit.p value="500"
iftrue="SELECTED"[/SP_MACRO]>500

NOTE: There are THREE changes per line where the values are edited. Also, a new line was added to the replacement – to display the choices 100 – 200– 300 – 400 – 500, rather than 50 – 100 – 150 – 200.

You DO NOT have to stop and restart WebSPIRS.

I have installed WebAdmin, but when I try to access it my server is unable to find the software.
In the /sproot/apache/conf/httpd.conf file, add the following entry:

ServerName IP

Where you replace "IP" with the IP address of the ERL server (or servername).

On the Solaris platform, the following (additional) information should be added:

Port 4416
User erl
Group erl
ServerName {IP or DNS}

Additional questions or problems should be directed to Technical Support.

Web Link to Holdings Vendor Configurations Guide
The following information may be used as a guide when configuring Web Link to Holdings within the webspirs_user.cfg file, located in webspirs/bin directory.

If your OPAC is not listed below, please contact Technical Support for further assistance.

These changes are made to the [SP.HOLDINGS] and the [VENDOR_1] sections. Make sure to set show.link.p=1, in each case, to turn on the "Check for holdings" button. Also, identify the vendor, and enter it in the entry:

vendor.p=XXX

Enter the same vendor information in the section heading below. The section heading would change from:

[VENDOR_1] to [XXX]

where XXX is the vendor identified in your "vendor.p=" entry

ALEPH
search.url.p=http://servername:4545/ALEPH/–/ext–find?base%3Diow01%26find%3D022%3D

link.graphic.p=cfh1.gif
issn.search.p=
isbn.search.p=

BIBSYS
search.url.p=http://servername/gate1/FIND?lang%3DE%26base%3DBIBSYS%26fo%3D%26F0%3D%26felt0%3Dto%26bd%3D%26F1%3D%26f elt1%3Dem%26F2%3D%26felt2%3Dde%26sn%3D
link.graphic.p=cfh1.gif
command.p=
search.suffix.p=%26F3%3D%26felt3%3Dok%26type%3DSearch
issn.search.p=
isbn.search.p=

Copac
search.url.p=http://servername/cgi–bin/cgiwrap/copacw/wzgw?terms%3D
link.graphic.p=cfh1.gif
command.p=
search.suffix.p=
issn.search.p=
isbn.search.p=

DRA
search.url.p=http://servername:port/XXX?key%3D
link.graphic.p=cfh1.gif
command.p=
search.suffix.p=%26ind%3DN
issn.search.p=
isbn.earch.p=

DRA – (another version)
search.url.p=http://servername/web2/tramp2.exe/do_authority_search/guest?setting_key%3Denglish%26servers%3D1home%26index%3D
link.graphic.p=cfh1.gif
command.p=
issn.search.p=)%26query%3D
isbn.search.p=(%26query%3D
INNOPAC
search.url.p=http://servername/search/i?SEARCH%3D
link.graphic.p=cfh1.gif
command.p=
search.suffix.p=
issn.search.p=
isbn.search.p=

PALS
search.url.p= http://servername/cgi–bin/pals–cgi?SET%20WEB%20NKLCAT____/MAXDI%201/is%20%20

PICA
search.url.p=http://servername/cgi–bin/nph–wwwredir/www.pica.nl:4901/IKTCOMMAND?ACTION%3DSEARCH%26IKT%3D8%26TERM%3D
link.graphic.p=cfh1.gif
command.p=
search.suffix.p=
issn.search.p=
isbn.search.p=

SIRSI – See information below

VOYAGER
search.url.p=http://servername/cgi–bin/XXX.cgi?DB%3Dlocal%26
link.graphic.p=cfh1.gif
command.p=
search.suffix.p=
issn.search.p=BOOL1%3Das%2Ba%2Bphrase%26FLD1%3DISSN%2B%28ISSN%29%26HIST%3D1%26CNT%3D10%2Brecords%2Bper%2Bpage%26SA1%3D
isbn.search.p=BOOL1%3Das%2Ba%2Bphrase%26FLD1%3DISBN%2B%28ISBN%29%26HIST%3D1%26CNT%3D10%2Brecords%2Bper%2Bpage%26SA1%

VOYAGER VER 99.1
search.url.p=http://servername/cgi–bin/Pwebrecon.cgi?v1%3D1%26CNT%3D10+records+displayed+per+page%26FT%3D

VOYAGER VER 2000
search.url.p=http://servername/cgi–bin/Pwebrecon.cgi?DB%3Dlocal%26
issn.search.p=BOOL1%3Das%2Ba%2Bphrase%26FLD1%3DISSN%2B%28ISSN%29%26CNT%3D10%26SAB1%3D
isbn.search.p=BOOL1%3Das%2Ba%2Bphrase%26FLD1%3DISBN%2B%28ISBN%29%26CNT%3D10%26SAB1%3D

VOYAGER 2000.1
search.url.p=http://servername/cgi–bin/Pwebrecon.cgi?DB%3Dlocal%26
link.graphic.p=cfh1.gif
command.p=
issn.search.p=BOOL1%3Das%2Ba%2Bphrase%26FLD1%3DISSN%2B%28ISSN%29%26CNT
%3D10%26SAB1%3D
isbn.search.p=BOOL1%3Das%2Ba%2Bphrase%26FLD1%3DISBN%2B%28ISBN%29%26CNT
%3D10%26SAB1%3D

VTLS
search.url.p=http://servername/cgi–bin/vtls40.web.gateway?searchtype%3Dcontrol+numcard%26searcharg%3D
link.graphic.p=cfh1.gif
command.p=
search.suffix.p=%26lang%3Denglish
issn.search.p=
isbn.search.p=

SIRSI:
Customers that have successfully linked to Webcat (versions 98.4 and later). However, to successfully implement the feature, cgi scripts for WebCat must be created. If you would like to follow the instructions provided by Sirsi, we will be able to configure WebSPIRS to link to that WebCat OPAC.

Below is some information that also was made available on the Sirsi List. This information is also available in the API Documentation from Sirsi. If the instructions are not sufficient, please call Sirsi Technical Support. They have been alerted that they may hear from SilverPlatter customers wanting to configure WebSPIRS for the Web Link to holdings.

Web Link to Holdings cannot be configured with earlier versions, prior to v. 98.4, of the SIRSI catalog.

STEP 1 (Create a cgi–bin script for WebCat)
Create a file called issn–search in the cgi–bin directory of your web server with the same ownership and permissions as the webcat cgi script.

The format of the issn–search file is described in the Unicorn API 98.4 Release Notes and is designed to support a search for ISSNs. Please note that the line beginning with "echo" and ending with ort_by}" needs to be entered as one continuous line without line breaks.

(contents of issn–search file)

#!/bin/sh
search_string=`echo "" | cut –d/ –f2`
unique_code=`date | sed 's/ /+/g'`
search_field="{022}"
kind_pol="GENERAL"
bind_pol="SUBJECT"

aths_pol="GENERAL"
spec_proc=""
label="words+or+phrase"
library="ALL"
pubyear=""
language="ANY"
format="ANY"
item_type="ANY"
location="ANY"
item_1cat="ANY"
item_2cat="ANY"
match_on="KEYWORD"
shadow="NO"
sort_by="NONE"

echo "Location: /uhtbin/cgisirsi/
/0/5?searchdata1=
&srchfield1=^^
^^&searchoper1=AND&thesaurus1=
&search_entries1=&search_type1=
&special_proc1=&library=
&pubyear=&language=&format=
&item_type=&location=&item1cat=$
{item1_cat}&item_2cat=&match_on=
&shadow=&sort_by="


echo

To do the same for ISBN search, you must create a file named isbn–search and set the search_field parameter equal to the marc tag for ISBN.

How do I change my default fields in WebSPIRS?
For WebSPIRS 3.1:

If you want to consistently display specific fields in your search results, such as the default list of fields (CITN) plus an additional field of your choice, i.e. the Descriptors Field (DE), you can make a change in the webspirs.cgi file to change the default fields to display. Edit the following line in that file:

/default.sp.nextform=top.htm

Change the line so that it will read:

/default.sp.nextform=top.htm\&sp.record.fields.brief.p=CITN,DE

For WebSPIRS 3.1 (NT):

Edit the following line in the webspirs.cmd:

cgibaby default.sp.nextform=top.htm

Change the line so that it will read:

cgibaby default.sp.nextform=top.htm sp.record.fields.brief.p=CITN,DE

For WebSPIRS 4.0:

If you would like to change the default fields in your search results in WebSPIRS 4.0, you would need to edit the /webspirs/bin/webspirs_user.cfg file. There are two places in the to set the default fields. They are:

[SP.RECORD]

fields.default.p

[SP.SHOW]

default.fields.p

Append the short name, in capital letters and separated by commas to the lines to reflect the additional fields you would like to add.

For Example, if you chose to display the descriptor field as default:

[SP.RECORD]

fields.default.p=CITN,*SHW,FTXT

would become:

fields.default.p=CITN,*SHW,FTXT,DE

and:

[SP.SHOW]

default.fields.p=CITN,*SHW,FTXT

would become:

default.fields.p=CITN,*SHW,FTXT,DE

How do I configure WebSPIRS 3.1 to print more than 50 records at a time?
In the WebSPIRS 3.1 interface, the maximum number of records listed in the drop–down list is 50. In order to change the number of records to be printed at one time, the pagesize.htm file must be modified. Pagesize.htm is located in the webspirs/template directory.

1. Back up a copy of the pagesize.htm by renaming it or copying it to another location before editing the file.

2. Edit the pagesize.htm file by adding lines for additional numbers of records options. A copy of pagesize.htm is shown below.

pagesize.htm

1
5
10
15
20
25
30
35
40
45
50

3. Save pagesize.htm, and test WebSPIRS.

I am using WebSPIRS 3.1 and Windows NT. How do I make configuration changes for automatically logging in, or for changing field defaults, etc.?
Edit the webspirs.cmd file (InetPub\scripts). There you designate the username, password, and server to access. You may set up a number of cmd files for various users. i.e. library1.cmd, library2.cmd, etc. You would then access these files by entering at the address line of the browser – http://(servername)/scripts/library1.cmd

Webspirs.cmd file would be edited to appear as follows:

@echo off
C:
cd "C:\Program Files\SilverPlatter\WebSPIRS\bin"
cgibaby default.sp.nextform=top.htm sp.username=(username) sp.password=(password) sp.record.fields.p=CITN,DE

How do I change the default number of records to be saved in WebSPIRS 3.1?
In order to change the number of records to be saved at one time, the pagesize.htm file must be modified. Pagesize.htm is located in the webspirs/template directory.

1. Back up a copy of the pagesize.htm by renaming it or copying it to another location before editing the file.

2. Edit the pagesize.htm file by adding lines for additional numbers of records options. A copy of pagesize.htm is shown below. 500 was added to the end of the file.

1
5
10
15
20
25
30
35
40
45
50
500

3. Save pagesize.htm, and test WebSPIRS.

How can configure WebSPIRS to select and entire database family at once?
In WebSPIRS 3.1:

The checking of the boxes on the database selection screen is controlled by the dblitmck.htm template.

Edit the dblitmck.htm template and change the end of this line:

iftrue="CHECKED" [/SP_MACRO]>

from "CHECKED" to "UNCHECKED" to switch the boxes from checked to unchecked. After making this change, the user is able to select and search databases, then click the "database" button and return to the database selection screen where no databases or checkboxes are selected.

In WebSPIRS 4:

WebSPIRS 4 has this functionality built in. Within the webspirs_user.cfg file, in the [SP.SHOW] section change the entry

db.menu.checkboxes.p=0

to:

db.menu.checkboxes.p=1

How can I turn off administrator e–mail messages in WebSPIRS 4.?
1. Edit the webspirs_user.cfg file in webspirs/bin directory.
2. In the [SP] section, remove your e–mail address from the line support.email.p=

You will no longer receive administrator e–mail messages.

How do I change the default number of records displayed in WebSPIRS 3.1? The default is 10 records at a time.
This involves editing a template in the webspirs/template directory. The template name is rechits.htm. Make a backup copy of the template before editing.

Edit the rechits.htm template and change the following four numeric values to the desired number of records. In the example below, the values have been changed to 50.

[SP_MACRO]sp.assign macro="sp.record.howmany.p" value="50" reset="yes"[/SP_MACRO]

WebSPIRS 4.x Web Link to Holdings with SIRSI catalog
Instructions for linking to the Unicorn v.98.4 OPAC are included in the API Documentation for that product. If the instructions are not sufficient please call Sirsi Technical Support. They have been alerted that they may hear from SilverPlatter customers wanting to configure WebSPIRS for Web Link to Holdings.

Web Link to Holdings cannot be configured with earlier versions, prior to v.98.4, of the SIRSI catalog.

How can I create a link on the WebSPIRS 4.x logout screen?
We suggest a change to the WebSPIRS logout.htm page, located in the file /home/webspirs/template/logout.htm. By editing this file as described below, a new link will display on the standard logout page, directing users to any designated web location:

Immediately following the line in the file which reads:

Your WebSPIRS session is now closed. Thank you for using WebSPIRS.
If you want to start a new session, click Start WebSPIRS

INSERT the following new line into the logout.htm file, substituting the appropriate URL for "http://www.silverplatter.com" and the appropriate text for "the SilverPlatter website"

I am receiving the following error when I try to access WebSPIRS 4.x
MISSING } IN COMPOUND STATEMENT

If you are using the Netscape browser, and your version is 3.01 or higher, you should check to see if you are using the unsupported 16–bit version by completing the following:

Open Netscape's Help menu from the toolbar and choose "About Plug–ins".

If the file name is NPNUL16.DLL, you are running the 16–bit version and you need to upgrade to the 32–bit version.

How can WebSPIRS 4.1 document delivery be configured for different user accounts?
The following steps show how this can be accomplished:

1. First provide a specialized WebSPIRS configurations for a user. This is conducted by using the 'spsubdir' command.

**NOTE**
For more information on 'spsubdir' and creating specialized user configurations, please visit this URL:

http://webspirs.silverplatter.com/ws40docs/6_bascfg.htm#diff defaults

2. Next, copy the entire contents of the /webspirs/bin/docdeliv directory to your user's sub directory.

3. Modify either ill.cfg, bl.cfg or cisti.cfg as appropriate.

**Example**

/webspirs/bin/test/docdeliv

The /docdeliv directory should contain /docdeliv/providers and /docdeliv/users

WebSPIRS 3.1 – errors caused by misconfiguration of the .htaccess file.
When WebSPIRS 3.1 is installed in conjunction with the Apache web server, access to directories is governed by the inclusion of IP addresses to the Apache .htaccess file. Adding entries to the file in an incorrect format can cause certain errors. For instance:
"Forbidden

You don't have permission to access /cgi–bin/ on this server."

This error is the result of your IP address not being included in the .htaccess file. The IP address should be listed in the form

"allow from [your_IP]"
for example:
"allow from 122.233.44.55"

If you wish to allow a certain range of IP addresses, list only the octets which are exclusive to the range you are filtering, such as

"allow from 123.234"

The above example will allow access to all IP addresses that begin with 123.234.

The following error is caused by incorrect syntax within the .htaccess file"

"Internal Server Error"

The server encountered an internal error or misconfiguration and was unable to complete your request.

Please contact the server administrator, root@localhost and inform them of the time the error occurred, and anything you might have done that may have caused the error."

If text entries are not in the correct format, the web server will not be able to interpret the file, causing the error. For example, note the third–from–last line in the example file below.

AuthUserFile /dev/null
AuthGroupFile /dev/null
AuthName InternalOnly
AuthType Basic

order deny,allow
deny from all
# main campus
allow from 121.55
# smith library
allow from 148.29.215
satellite campus
allow from 190.172.133

Entering the "satellite campus" text into the file with no comment mark (#) preceding it will cause the Internal Server Error message. Editing the file and including the comment mark will resolve the error.

In WebSPIRS 4, how do I set my default field set to show all fields?
Open the webspirs_user.cfg file in the /webspirs/bin directory.
Edit the following in the [SP.SHOW] section:
default.fields.p=*A

All fields will be displayed within the database.

Quick guide to create configuration section for SilverPlatter's WebLink to Holdings feature:
1. Access your OPAC from a web browser and search for an ISSN number which you know is in your OPAC database.

2. Examine the location toolbar of the browser and identify the position of ISSN number in the URL.

3. Select the part of URL preceding the ISSN number, paste it into the search.url.p= parameter of appropriate vendor section in the webspirs_user.cfg file (for more information on the webspirs_user.cfg consult WebSpirs Implementor's Guide)

4. Select the part of URL following the ISSN number and paste it into the search.suffix.p= parameter in webspirs_user.cfg file.

5. Replace all "=" and "&" characters in these two parts of URL with their respective hex values "%3D" and "%26".

6. Ensure that the vendor.p= parameter in webspirs_user.cfg, section [SP.HOLDINGS] points to appropriate vendor section.

7. Ensure that the parameter show.link.p=1 and, if you want to limit the weblink to a subset of your SilverPlatter databases, that you have listed appropriate database SET–IDs in link.databases.p=

How to activate and use the ILL Document Delivery Service in WebSPIRS 4.*
Currently WebSPIRS 4.* comes with three pre–configured email based Document Delivery Services. It allows you to send an email request with information from the bibliographic record, to existing document delivery suppliers (British Library and CISTI) or to a email address of your choice e.g. a Inter Libary Loan emailaddress. These services (all 3) are activated at the same by editing the [DD]–section of the file ../webspirs/bin/webspirs_user.cfg file. In the [DD] section you just need to activate (i.e. un–comment) the following line:
root.spec.directory.p=./docdeliv

Once enabled, an "ORDER" button will show with the bibliographic record. Which of the services show for the database in use depends on the configuration files that are located in ../webspirs/bin/docdeliv/providers/. They are called "bl.cfg", "cisti.cfg" and "ill.cfg". The latter is the configuration file for a customer–specific Document Delivery service. The eMail address of the ILL service is put in the "destination.p=...." line of the ill.cfg file.

If you do not need one or more of these services you can simply disable them by removing the corresponding .cfg files from the ../docdeliv/providers directory and they will no longer appear as available service.

If you want to have your favourite delivery mode (fax, interoffice mail, courier) on top of the list of your ILL request form you just have to change the order of the "method" parameters in the[dd.delivery] section in ill.cfg.

In ill.cfg there is also a parameter which controls the text of the delivery mode label in the form: e.g. method1.name.p="Interoffice mail", and the name of the corresponding template files that are shown as order forms in WebSPIRS are: illfax.htm, illmail.htm, illcourier.htm.

Changing the WebSPIRS 4.3 appearance using the open.location.p setting
There are 3 modes available for configuration in the webspirs_user.cfg file, located in the /webspirs/bin directory. They may be summarized as follows:

Open.location.p=0

This setting simulates the WebSPIRS 4.2 – with the "Bring it back" or "Session in Progress" page behind the open window.

Open.location.p=1

This setting allows you to use WebSPIRS within a browser window. All broser buttons are fully functioning.

Open.location.p=2

(default) This setting opens WebSPIRS in an independent window. There are no browser buttons visible.

SilverPlatter CAUTIONS customers who choose to set the mode to "open.location.p=1". The browser buttons are visible. We suggest users NOT to use the "BACK" browser button. Using the browser "BACK" button can result in unexpected results.

I receive the following error, when I try to login to WebSPIRS:
"Maximum number of users has temporarily been reached."

This error occurs when the number of users trying to access WebSPIRS exceeds the number allowed for the purchased subscription. The suggestion is to wait a length of time and try again later. If you see this error repeatedly, please notify your librarian.

"Access Denied" When You Try to Start the WebSPIRS v4.x NT Service
If you get an "Access Denied" message when you try to start the WebSPIRS v4.x NT Service, the server cannot locate webspirsd.exe.

The Windows NT Service Manager truncates directory names to 8.3 format. This causes a conflict with the default WebSPIRS NT installation path: (C:\ProgramFiles\SilverPlatterInformation\WebSPIRS\bin\webspirsd.exe), especially if you have WebSPIRS 3.1 installed in it's default directory (C:\ProgramFiles\SilverPlatter\WebSPIRS) on the same machine. Installing WebSPIRS v4.x to another directory should resolve the problem.

Trouble Starting, Restarting or Stopping WebSPIRS v4.x (UNIX)
If you have trouble starting, restarting or stopping WebSPIRS, there may be more than one ps command on your system or the command may not be in your path.

1. Open the webspirs script in the webspirs/bin directory and find the following line:

PSCMD="ps –ef"

2. Edit the line to include the exact path to the ps command. For example:
if the command is in the /bin
directory:

PSCMD="/bin/ps –ef"

3. Save and close the file.

JavaScript Error: http://www.your_server.com:8000/login, line 1168: missing } in compound statement.
If you get the above error message when you try to open WebSPIRS with Netscape Navigator, you may be using the unsupported 16–bit version.

To find out what version you are using:

Open Netscape Navigator and select About Plug–ins from the HELP menu.

If the file name is NPNUL16.DLL, you are running the 16–bit version of Netscape Navigator. In order to use WebSPIRS 4.1, you must upgrade to one of the WebSPIRS v4.x supported web browsers.

Pages Not Loading Completely/JavaScript Errors Because of Bad Modem Connection
A bad modem connection could prevent WebSPIRS v4.x from loading pages completely and result in JavaScript errors. If your users connect to WebSPIRS via modem, you may want to configure WebSPIRS to bypass the Login page (as it tends to take longer to load) and auto–open specific databases.

NOTE: Bypassing the Login and Database Selection pages may reduce the occurrences of this problem for modem users, but it may not eliminate the problem completely.

"Failed to bind on port# (number)" when starting WebSPIRS v4.x
If a "failed to bind on port# (number)" message is displayed when you start WebSPIRS (webspirsd), there is a port number conflict. In UNIX, use the process command to see what processes are running. One of the following conditions may have occurred:

WebSPIRS is already running and using that port number. You can try stopping and restarting WebSPIRS to clear up the conflict.

In some cases, you may find that a port number is hung, but no process is using it. In this case, wait for 15 to 20 minutes for the port to time out and then try starting WebSPIRS.

Another process is using the port number. In this case, you can change the port number used by WebSPIRS:

1. Change to the webspirs/bin directory and open the webspirsd.cfg file in a text editor.

2. Locate these lines in the file:

port=8590
request_port=8591
request_name=:8591 (Note colon.)

3. Change the port number to another number (we recommend a number over 8000).

4. Change request_port and request_name to a number that is one greater than the port number.

NOTE: Both request_port and request_name should have the same number.

5. Save and close the file.

6. Restart WebSPIRS.

Complete WebSPIRS v4.x Page Not Displayed (Internet Explorer Only)
For Internet Explorer users and in some conditions, portions of a WebSPIRS page may not load. To display the complete page you can press a button in the toolbar to go to another page of the interface. Then, return to the page on which the problem occurred. The full page should then be displayed.

Pressing WebSPIRS v4.x Login Button, No Response
If a user gets no response when they click on the Login button, ensure that JavaScript is enabled in the browser.

Table of Contents Not Displayed for a Full–Text Database
If a user opens the Table of Contents while searching a full–text database and the Table of Contents does not display, ensure that Java is enabled in the browser. The Table of Contents for Full–Text databases requires the use of Java.

Shutdown failed on socket 4 error 131 – WebSPIRS v4.03, 4.1b2
This error message appears only in the window (console) of the UNIX machine where WebSPIRS is installed whenever a user logs into WebSPIRS.

This message only appears when the WebSPIRS software has been started by ./webspirs start or ./webspirs restart. When WebSPIRS is automatically started by the rebooting the server the shutdown failed message does not appear. This message does not affect searching.

The socket number in the message may vary.

This error message is harmless and may be ignored. It will not appear in the next release of WebSPIRS, v. 4.1.

WebSPIRS 3.1 reports "connect to adult failed"
This error may be caused when there is a problem with WebSPIRS reading the ERL DBServer dbmenu.cfg file. This file is used for a custom listing of the databases.

Open the webspirs.cfg file, found in /webspirs/bin and look for the line

StructuredDBlist=.

If it is = to 1 then it is looking for the dbmenu.cfg file

If it is = to 0 then it displays the databases list in the order the databases were installed.

In order to allow WebSPIRS to display the databases while you are editing the dbmenu.cfg file set the StructuredDBlist=0.

For more information on the dbmenu.cfg file click here

"An error has occurred. There was a problem processing your last request. Reload this page. If you see this page again after reloading a few times, contact your administrator."
This error occurs when attempting to connect to WebSPIRS and is typically due to omitting the question mark at the end of the URL. WebSPIRS 4+ requires the URL to have a question mark at the end.
(spweb.silverplatter.com/erl2?)

Try the URL again and add the question mark to the end of the URL. If this message still appears, please contact the Network Administrator of your institution or support@silverplatter.com. Please include the problem URL in the email.

Suggestions for correcting problems making WebSPIRS 3.1 work with Apache.
When installing WebSPIRS 3.1 with the Apache 1.3 web server you may run into certain problems. Although there may be several ways of making WebSPIRS 3.1 work with the Apache web server, the following example may assist you. Please note that you may have Apache and WebSPIRS installed in different locations as in this example. For further information please use the WebSPIRS 3.1 and Apache documentation.
1. install WebSPIRS in /sproot/webspirs31
2. as "htdocs" directory choose /sproot/webspirs31
3. as "cgi–bin" directory choose /usr/local/apache/cgi–bin.
4. before (re)starting the apache server add two lines to httpd.conf:

ScriptAlias /cgiscripts /sproot/webspirs31/bin/
Alias /webspirs /sproot/webspirs31/url

5. edit /sproot/webspirs31/url/webspirs.htm and change the link to acces webspirs from ACTION="/cgi–bin/webspirs.cgi" to ACTION="/cgiscripts/webspirs.cgi"

How to run multiple WebSPIRS 4.* on the same machine
In principle one can run multiple WebSPIRS4+ daemons on the same machine, provided they run on different ports. For each of the supported platforms, NT, Solaris, Linux and AIX*. This is in fact relatively easy. Simply copy the WebSPIRS file structure to a new location and edit the webspirsd.cfg file located in the bin directory. For instance if you want to run WebSPIRS on port 8000 you should use the following values:

port=8000
request_port=8001
request_name=:8001

To start this "copy of WebSPIRS" from the command line you go to ...../webspirs/bin and enter:
– On Unix platdorms: ./webspirs start
– On the NT platform: webspirsd.exe –console

If you install a new version of WebSPIRS on a machine that is already running WebSPIRS then this is also possible, and again a bit different for NT and UNIX.

UNIX:
On the UNIX platforms (Linux, Solaris and AIX*) you can install a new version of WebSPIRS in a different location. During the installation you are asked to specify a port number and you must ensure this number is different from the port used by the existing WebSPIRS daemon. When the install script asks you if you want to start the (new) WebSPIRS automatically, you should be aware that this will update the SM0webspirs file in your systems /etc/rc.... directories. Should you require to change a start–up file, or add additional start–up files for the multiple versions of WebSPIRS, then is possible. Please contact SilverPlatter Technical support if you need assistance with this: support@silverplatter.com

NT:
On NT you can also install a new version of WebSPIRS without un–installing the current version of WebSPIRS on the machine. You can even leave the current WebSPIRSd service running and start the install of the new version. In this example you install WebSPIRS 4.2 while version 4.11 is already installed and running

1) start the WebSPIRS 4.2 installation
2) make sure you install WebSPIRS 4.2 in a different location
3) When it detects the already running WebSPIRS service the set–up process will prompt you, and you can say "no, do not stop the running service"
4) Complete the install process and do not forget to run WebSPIRS on a different port number then the already existing WebSPIRS.

At this point you can start WebSPIRS 4.2 ether from the command line as a service that will also start automatically when the machine is restarted. In any case the install of WebSPIRS 4.2 will change the service setting in the registries to the location of the WebSPIRS 4.2 daemon. In case you want change this and prefer another WebSPIRS to run as the service you need to change the following registry entries to point to the correct locations:
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\AppPaths\webspirsd.exe
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\WebSPIRSd
Please contact SilverPlatter Technical support if you need assistance with this: support@silverplatter.com *Note that for AIX, WebSPIRS 4.11 was the last version. Later versions of WebSPIRS will not be made available on AIX.

How to run WebSPIRS 4.* on port 80.
WebSPIRS 4.* contains it's own build in http daemon. The default port number for the WebSPIRS daemon is 8590, but you can easily change this at any time. If you like to run WebSPIRS on port 80, which is the default port number for http then this is possible. Before you make this change please verify if there is not already another webserver on running your server using port 80.. Once you are sure this is not the case, then edit the webspirsd.cfg file located in the WebSPIRS bin directory and restart WebSPIRS.
To run WebSPIRS on port 80 you should use the following values in webspirsd.cfg:

port=80
request_port=81
request_name=:81

Please note that for any UNIX system, if you want an application running on any port number below 1024, you need to run it with root permissions. As a consequence you need to edit the SM0webspirs start–up file located in /etc/rc2.d (differs per UNIX flavor) make sure WebSPIRS starts on port 80, e.g. WSUSER=root.

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.


WINSPIRS 4.01

General

Configuration Options

Error Messages

Installation


When attempting to access network databases, WinSPIRS freezes and the text "initializing" displays.
During initialization, WinSPIRS searches all network drives looking for SilverPlatter databases. To do this, it searches and attempts to open files with an extension of .dir or .21. Typically, when WinSPIRS freeze at the "initializing" prompt, it has found a .dir or .21 file it is unable to open.

You can determine which network drive or file is causing WinSPIRS to freeze by unmapping the various network drives, then remapping them one at a time to see which drive may be causing WinSPIRS to freeze. When the drive is identified, you can do a file search on this drive to determine what files may be present that have an extension of .dir or .21. In one instance, the customer's e–mail system was creating files of this type and causing WinSPIRS to freeze. The resolution was to map the networked CDs to a different network drive.

Once you have determined which drive may be conflicting, move the SilverPlatter databases to a different network drive or configure WinSPIRS for subdirectory access and move the file to a subdirectory.

How do I get WinSPIRS to work with WinU? When I launch the WinSPIRS program through the WinU security interface, it quits after about 15 seconds.
WinSPIRS can be launched through WinU if a 3–5 second launch pause is configured for the WinSPIRS button in WinU. Editing the WinSPIRS profile from within the WinU administrative interface allows the pause to be configured.

How do I change the default Show/Print/Save fields in WinSPIRS?
Run the WinSPIRS reconfiguration program, select Workstation Options. Click on the box marked "allow setting of default fields". Save and exit the Reconfig program. When WinSPIRS is next started, a button entitled "Make Default Fields" will be enabled under the Options section where fields are selected. Clicking on the button after changing the field display will update the default fields. To keep the new default fields set, run the reconfiguration program and deselect the checkbox to disable the button.

How do I configure WinSPIRS to map a drive letter to a database only when a users launches WinSPIRS and unmaps the drive letter when the user exits WinSPIRS?
Mapping a drive when needed and then unmapping the drive when the operation is complete, is referred to as "dynamic mapping". WinSPIRS may be configured to map drive letters for volumes on Novell NetWare, Windows for Workgroups, Windows 95, or Windows NT servers. Drive mappings are created when starting WinSPIRS and deleted when ending WinSPIRS. For Novell NetWare volumes, the user must log into the Novell server prior to starting WinSPIRS.

WinSPIRS only uses available (unmapped) drive letters starting from D:. The software assumes that the C: drive already exists. It attempts to create a mapping for each path read from the specified mapping section in the WINSPIRS.INI file. First, create a new section in the WINSPIRS.INI file that has the following form:

Novell NetWare:

[MAPPING_SECTION_NAME]
MAP1=SERVER/VOLUME:\
MAP2=SERVER/VOLUME:\

WFWG, Windows 95, and Windows NT:

[MAPPING_SECTION_NAME]
MAP1=\\SERVER\VOLUME\
MAP2=\\SERVER\VOLUME\

For example:

Novell NetWare:

[MEDLINE]
MAP1=SUPPORT/MXA5:\
MAP2=SUPPORT/MX5A:\
MAP3=SUPPORT/MX5B:\

[CINAHL]
MAP1=SUPPORT/NUC5:\

WFWG, Windows 95, and Windows NT:

[MEDLINE]
MAP1=\\SUPPORT\MXA5\
MAP2=\\SUPPORT\MX5A\
MAP3=\\SUPPORT\MX5B\

[CINAHL]
MAP1=\\SUPPORT\NUC5\

These sections should be configured by the administrator and can be given any name the administrator wishes. However, they must contain entries of the form Map1, Map2, ... etc. which specify the network server name and volume name.

To create the drive mappings, the relevant mapping section should be passed as an argument to WINSPIRS.EXE in the Command Line section of the icon's properties. The command line has the following form:

WINSPIRS.EXE –MMAPPING_SECTION_NAME (The first 'M' is necessary.)

For example:

C:\WINSPIRS\WINSPIRS.EXE –MCINAHL

You can then create multiple WinSPIRS icons to start different mapping sections if desired.

Setting default fields in WinSPIRS
Administrators can specify the default field set for showing, printing, and downloading records in WinSPIRS. The field sets are saved on a per database basis. Below are the instructions for setting defaults:

1. Start the WinSPIRS Reconfiguration program by double–clicking on its icon.
2. Click on the WORKSTATION OPTIONS button.
3. Click on ALLOW SETTING OF DEFAULT FIELDS to place an X in the box.
4. Click the OK button, then click the SAVE button.
5. Start WinSPIRS and open the appropriate database. If you wish to configure the same field defaults for all SilverPlatter databases, open all databases.
6. From the Options menu, select Print Options, Show Options, or Download Options.
7. Select the desired fields and click on the MAKE DEFAULT FIELDS button.
8. Repeat steps 6–7 to set the default fields for the remaining Options.
9. Since default fields must be configured separately for each SilverPlatter database, repeat steps 5–8 for other databases.
Exit WinSPIRS and repeat steps 1–4 to clear the X from the ALLOW SETTING OF DEFAULT FIELDS box. This will remove the MAKE DEFAULT FIELDS button from WinSPIRS to prevent unauthorized changes to the field sets.

MEDLINE EXPRESS SUBDIRECTORY MAPPING WITH WinSPIRS
WinSPIRS includes the capability to check subdirectories for SilverPlatter databases. This means that only one drive letter is required when mapping compact discs. The following instructions will configure WinSPIRS for this type of access:

1. Map a drive letter to the volume containing SilverPlatter databases.
2. Make note of each subdirectory that contains a SilverPlatter disc.
3. Change to the WinSPIRS directory and edit the WINSPIRS.INI file using a DOS text editor.
4. Move to the section entitled [CORE] in the WINSPIRS.INI file.
5. Add the path of each disc's subdirectory to the DRIVES parameter and separate them with a semicolon. For example, if drive G: is mapped to the compact discs and the subdirectory names matched each CD's volume label, the line might be:

DRIVES=G:\MX19;G:\MX9A;G:\MX9B;G:\MX9C;G:\MX9D;G:\NU07;G:\ER03

6. Ensure that the [CORE] section of WINSPIRS.INI also contains the parameter:

MAGDRIVE=1

Once this is done, WinSPIRS will check each subdirectory in the list for a SilverPlatter database.

However, when using this configuration with MEDLINE Express, additional steps must be taken. Each set of MEDLINE Express has a *.DIR file that tells the software where to find specific files, using each disc's volume label as a path. These paths must be changed so that they appear as subdirectories instead of volumes. Disc 1, Disc 14 and Disc 15 will work without making any changed to their *.DIR files. The following instructions can be used for Discs 2–4, Discs 5–7, Discs 8–10, Discs 11–13 and Discs 16–19:

1. Copy the MX9A.DIR file from MEDLINE Express Disc 2 to the \WINSPIRS\DBASE subdirectory.
2. Edit the MX9A.DIR file and replace each volume name (MX9A, MX9B, MX9C) with the appropriate drive letter and subdirectory name. For example, if Disc 2 is located in the G:\MX9A directory and Disc 3 is in the G:\MX9B directory and Disc 4 is in the G:\MX9C directory, then the edited file would be as follows:

DISC.dbl =MX9A.dbl

MX9Z.frq = +G:\MX9A\\medx.frq
MX9Z.d = +G:\MX9A\\medx.ftd
MX9Z.v = +G:\MX9A\\medx.ltd
MX9Z.i0 = +G:\MX9A\\medx1.inv+G:\MX9B\\medx2.inv
MX9Z.t0 = +G:\MX9B\\medx1.ctx+G:\MX9C\\medx2.ctx
MX9Z.x = +G:\MX9A\\medx.cti
MX9Z.tti = +G:\MX9A\\medx.tti

MX9Z.f = +G:\MX9A\\medx.ftr
MX9Z.e = +G:\MX9A\\medx.ent
MX9Z.mi = +G:\MX9A\\medx.gi
MX9Z.mf = +G:\MX9A\\medx.gf
MX9Z.mt = +G:\MX9A\\medx.gt
MX9Z.ms = +G:\MX9A\\medx.gs
MX9Z.me = +G:\MX9A\\medx.ge
MX9Z.men =mesh90.men

THES.ver = +G:\MX9A\\mesh99.ver
MESH.f90 = +G:\MX9A\\mesh.ftr
MESH.sbh = +G:\MX9A\\mesh.sbh
MESH.txt = +G:\MX9A\\mesh.txt
MESH.ptd = +G:\MX9A\\mesh.ptd
MESH.ptt = +G:\MX9A\\mesh.ptt
MESH.mtd = +G:\MX9A\\mesh.mtd
MESH.tnd = +G:\MX9A\\mesh.tnd
MESH.tnt = +G:\MX9A\\mesh.tnt
MESH.doc =meshhelp.doc
MESH.err =mesh.err

DISC.ovw = +G:\MX9A\\disc.ovw
DISC.scr =disc.scr
DISC.dcs =disc.dcs
DISC.err =disc.err
ISIS.men =disc.men
NHLP.doc =nhlp.doc /* new help index file
CHLP.doc =chlp.doc /* new help context file

MX9Z.fam =MEDX
MEDX.cdt =MEDX.CDT
MX9Z.gen =MEDX.GEN
MX9Z.hlp =MEDX.HLP

MX9A.nam =MEDLINE Express Disc 2
MX9B.nam =MEDLINE Express Disc 3
MX9C.nam =MEDLINE Express Disc 4

3. Copy the MX9D.DIR file from MEDLINE Express Discs 5 to the \WINSPIRS\DBASE subdirectory.
4. Edit the MX9D.DIR file and replace each volume name (MX9D, MX9E, MX9F) with the appropriate drive letter and subdirectory name. For example, if Disc 5 is located in the G:\MX9D directory and Disc 6 is in the G:\MX9E directory and Disc 7 is in the G:\MX9F directory, then the edited file would be as follows:

DISC.dbl =MX9D.dbl

MX9Y.frq = +G:\MX9D\\medx.frq
MX9Y.d = +G:\MX9D\\medx.ftd
MX9Y.v = +G:\MX9D\\medx.ltd
MX9Y.i0 = +G:\MX9D\\medx1.inv+G:\MX9E\\medx2.inv
MX9Y.t0 = +G:\MX9E\\medx1.ctx+G:\MX9F\\medx2.ctx
MX9Y.x = +G:\MX9D\\medx.cti
MX9Y.tti = +G:\MX9D\\medx.tti

MX9Y.f = +G:\MX9D\\medx.ftr
MX9Y.e = +G:\MX9D\\medx.ent
MX9Y.mi = +G:\MX9D\\medx.gi
MX9Y.mf = +G:\MX9D\\medx.gf
MX9Y.mt = +G:\MX9D\\medx.gt
MX9Y.ms = +G:\MX9D\\medx.gs
MX9Y.me = +G:\MX9D\\medx.ge
MX9Y.men =mesh90.men

THES.ver = +G:\MX9D\\mesh99.ver
MESH.f90 = +G:\MX9D\\mesh.ftr
MESH.sbh = +G:\MX9D\\mesh.sbh
MESH.txt = +G:\MX9D\\mesh.txt
MESH.ptd = +G:\MX9D\\mesh.ptd
MESH.ptt = +G:\MX9D\\mesh.ptt
MESH.mtd = +G:\MX9D\\mesh.mtd
MESH.tnd = +G:\MX9D\\mesh.tnd
MESH.tnt = +G:\MX9D\\mesh.tnt
MESH.doc =meshhelp.doc
MESH.err =mesh.err

DISC.ovw = +G:\MX9D\\disc.ovw
DISC.scr =disc.scr
DISC.dcs =disc.dcs
DISC.err =disc.err
ISIS.men =disc.men
NHLP.doc =nhlp.doc /* new help index file
CHLP.doc =chlp.doc /* new help context file

MX9Y.fam =MEDX
MEDX.cdt =MEDX.CDT
MX9Y.gen =MEDX.GEN
MX9Y.hlp =MEDX.HLP

MX9D.nam =MEDLINE Express Disc 5
MX9E.nam =MEDLINE Express Disc 6
MX9F.nam =MEDLINE Express Disc 7

5. Copy the MX9G.DIR file from MEDLINE Express Disc 8 to the \WINSPIRS\DBASE subdirectory.
6. Edit the MX9G.DIR file and replace each volume name (MX9G, MX9H, MX9I) with the appropriate drive letter and subdirectory name. For example, if Disc 8 is located in the G:\MX9G directory and Disc 9 is in the G:\MX9H directory and Disc 10 is in the G:\MX9I directory, then the edited file would be as follows:

DISC.dbl =MX9G.dbl

MX9X.frq = +G:\MX9G\\medx.frq
MX9X.d = +G:\MX9G\\medx.ftd
MX9X.v = +G:\MX9G\\medx.ltd
MX9X.i0 = +G:\MX9G\\medx1.inv+G:\MX9H\\medx2.inv
MX9X.t0 = +G:\MX9H\\medx1.ctx+G:\MX9I\\medx2.ctx
MX9X.x = +G:\MX9G\\medx.cti
MX9X.tti = +G:\MX9G\\medx.tti

MX9X.f = +G:\MX9G\\medx.ftr
MX9X.e = +G:\MX9G\\medx.ent
MX9X.mi = +G:\MX9G\\medx.gi
MX9X.mf = +G:\MX9G\\medx.gf
MX9X.mt = +G:\MX9G\\medx.gt
MX9X.ms = +G:\MX9G\\medx.gs
MX9X.me = +G:\MX9G\\medx.ge
MX9X.men =mesh90.men

THES.ver = +G:\MX9G\\mesh99.ver
MESH.f90 = +G:\MX9G\\mesh.ftr
MESH.sbh = +G:\MX9G\\mesh.sbh
MESH.txt = +G:\MX9G\\mesh.txt
MESH.ptd = +G:\MX9G\\mesh.ptd
MESH.ptt = +G:\MX9G\\mesh.ptt
MESH.mtd = +G:\MX9G\\mesh.mtd
MESH.tnd = +G:\MX9G\\mesh.tnd
MESH.tnt = +G:\MX9G\\mesh.tnt
MESH.doc =meshhelp.doc
MESH.err =mesh.err

DISC.ovw = +G:\MX9G\\disc.ovw
DISC.scr =disc.scr
DISC.dcs =disc.dcs
DISC.err =disc.err
ISIS.men =disc.men
NHLP.doc =nhlp.doc /* new help index file
CHLP.doc =chlp.doc /* new help context file

MX9X.fam =MEDX
MEDX.cdt =MEDX.CDT
MX9X.gen =MEDX.GEN
MX9X.hlp =MEDX.HLP

MX9G.nam =MEDLINE Express Disc 8
MX9H.nam =MEDLINE Express Disc 9
MX9I.nam =MEDLINE Express Disc 10

7. Copy the MX9J.DIR file from the MEDLINE Express Disc 11 to the \WINSPIRS\DBASE subdirectory.
8. Edit the MX9J.DIR file and replace each volume name (MX9J, MX9K, MX9L) with the appropriate drive letter and subdirectory name. For example, if Disc 11 is located in the G:\MX9J directory and Disc 12 is in the G:\MX9K directory and Disc 13 is in the G:\MX9L directory, then the edited file would be as follows:

DISC.dbl =MX9J.dbl

MX9W.frq = +G:\MX9J\\medx.frq
MX9W.d = +G:\MX9J\\medx.ftd
MX9W.v = +G:\MX9J\\medx.ltd
MX9W.i0 = +G:\MX9J\\medx1.inv+G:\MX9K\\medx2.inv
MX9W.t0 = +G:\MX9K\\medx1.ctx+G:\MX9L\\medx2.ctx
MX9W.x = +G:\MX9J\\medx.cti
MX9W.tti = +G:\MX9J\\medx.tti

MX9W.f = +G:\MX9J\\medx.ftr
MX9W.e = +G:\MX9J\\medx.ent
MX9W.mi = +G:\MX9J\\medx.gi
MX9W.mf = +G:\MX9J\\medx.gf
MX9W.mt = +G:\MX9J\\medx.gt
MX9W.ms = +G:\MX9J\\medx.gs
MX9W.me = +G:\MX9J\\medx.ge
MX9W.men =mesh90.men

THES.ver = +G:\MX9J\\mesh99.ver
MESH.f90 = +G:\MX9J\\mesh.ftr
MESH.sbh = +G:\MX9J\\mesh.sbh
MESH.txt = +G:\MX9J\\mesh.txt
MESH.ptd = +G:\MX9J\\mesh.ptd
MESH.ptt = +G:\MX9J\\mesh.ptt
MESH.mtd = +G:\MX9J\\mesh.mtd
MESH.tnd = +G:\MX9J\\mesh.tnd
MESH.tnt = +G:\MX9J\\mesh.tnt
MESH.doc =meshhelp.doc
MESH.err =mesh.err

DISC.ovw = +G:\MX9J\\disc.ovw
DISC.scr =disc.scr
DISC.dcs =disc.dcs
DISC.err =disc.err
ISIS.men =disc.men
NHLP.doc =nhlp.doc /* new help index file
CHLP.doc =chlp.doc /* new help context file

MX9W.fam =MEDX
MEDX.cdt =MEDX.CDT
MX9W.gen =MEDX.GEN
MX9W.hlp =MEDX.HLP

MX9J.nam =MEDLINE Express Disc 10
MX9K.nam =MEDLINE Express Disc 11
MX9L.nam =MEDLINE Express Disc 12

9. Copy the MX9O.DIR file from MEDLINE Express Disc 16 to the \WINSPIRS\DBASE subdirectory.
10. Edit the MX9O.DIR file and replace each volume name (MX9O, MX9P, MX9Q, MX9R) with the appropriate drive letter and subdirectory name. For example, if Disc 16 is located in the G:\MX9O directory and Disc 17 is in the G:\MX9P directory and Disc 18 is in the G:\MX9Q directory and Disc 19 is in the G:\MX9R directory, then the edited file would be as follows:

DISC.dbl =MX9O.dbl

MX9V.frq = +G:\MX9O\\medx.frq
MX9V.d = +G:\MX9O\\medx.ftd
MX9V.v = +G:\MX9O\\medx.ltd
MX9V.i0 = +G:\MX9O\\medx1.inv+G:\MX9P\\medx2.inv
MX9V.t0 = +G:\MX9P\\medx1.ctx+G:\MX9Q\\medx2.ctx+G:\MX9R\\medx3.ctx
MX9V.x = +G:\MX9O\\medx.cti
MX9V.tti = +G:\MX9O\\medx.tti

MX9V.f = +G:\MX9O\\medx.ftr
MX9V.e = +G:\MX9O\\medx.ent
MX9V.mi = +G:\MX9O\\medx.gi
MX9V.mf = +G:\MX9O\\medx.gf
MX9V.mt = +G:\MX9O\\medx.gt
MX9V.ms = +G:\MX9O\\medx.gs
MX9V.me = +G:\MX9O\\medx.ge
MX9V.men =mesh90.men

THES.ver = +G:\MX9O\\mesh99.ver
MESH.f90 = +G:\MX9O\\mesh.ftr
MESH.sbh = +G:\MX9O\\mesh.sbh
MESH.txt = +G:\MX9O\\mesh.txt
MESH.ptd = +G:\MX9O\\mesh.ptd
MESH.ptt = +G:\MX9O\\mesh.ptt
MESH.mtd = +G:\MX9O\\mesh.mtd
MESH.tnd = +G:\MX9O\\mesh.tnd
MESH.tnt = +G:\MX9O\\mesh.tnt
MESH.doc =meshhelp.doc
MESH.err =mesh.err

DISC.ovw = +G:\MX9O\\disc.ovw
DISC.scr =disc.scr
DISC.dcs =disc.dcs
DISC.err =disc.err
ISIS.men =disc.men
NHLP.doc =nhlp.doc /* new help index file
CHLP.doc =chlp.doc /* new help context file

MX9V.fam =MEDX
MEDX.cdt =MEDX.CDT
MX9V.gen =MEDX.GEN
MX9V.hlp =MEDX.HLP

MX9O.nam =MEDLINE Express Disc 16
MX9P.nam =MEDLINE Express Disc 17
MX9Q.nam =MEDLINE Express Disc 18
MX9R.nam =MEDLINE Express Disc 19

How do I bypass the database selection screen and automatically open a database using WinSPIRS?
I. Create a new icon and change the WinSPIRS command line

For example, to automatically open 5 CAB databases:

1. Create an icon on your desktop to run WinSPIRS. (You may choose to create multiple icons, name them according to the database to which you want to jump, and configure each as follows.) Right click on the icon and choose 'properties'.
2. Choose the 'shortcut' tab. On the target line add ' CI??', so that the end of the line appears as:

WINSPIRS.EXE CI??

This will automatically load all of the CAB databases which have the database ID beginning with 'CI' and contain a 4–digit database ID. (If you do not know the ID for the database you wish to configure, contact Technical Support.)

II. To jumpstart directly to the search area when connecting to an ERL server, you need to automatically login to your server. You configure autologin as follows:

1. Edit the ERLCLNT.CFG file in the WinSPIRS directory and set:

server_addr1=/2/(the name of the ERL server to which you want to connect)/416

Please note that there should only be one SERVER_ADDR1 parameter in the file. Any other occurrences of this parameter should begin with /*

(e.g., /* server_addr1 = /2/erl3.silverplatter.com/416).

2. In the WinSPIRS group under Program Manager, double–click on the WinSPIRS RECONFIGURATION icon.
3. Within the Configuration Options menu, click once on ERL OPTIONS.
4. In the Mode section of the ERL Options dialog box, select either ERL CLIENT ONLY or LOCAL AND ERL mode. The second option will access databases on both the local workstation and on any configured ERL servers.
5. In the Automatic Login section, click once on the ADD button.
6. Within the New Server dialog box, add the Server ID, Username and Password:

Server ID (blank for default):
Username: (your username)
Password: (your password)
7. Click once on the OK button to accept the new settings.
8. Finally, click once on the OK button to accept the new ERL options and once on the SAVE button to save the settings.

II. To jump past the START screen and open directly to the search screen:
1. Open the Reconfig Program of WinSPIRS
2. Choose 'Workstation Options'. Remove the check within the box marked 'Display Start Button'. Press OK to than screen, then Save.

General configuration instructions for a networked version of WinSPIRS 4.0.
If you wish to set up one WinSPIRS program to be shared by many users across the network, choose a shared drive letter to locate the file. You can download win40.exe at a workstation and during installation, choose where to place the file. During installation, choose "shared" and "shared", when given options. Follow the instructions during installation and place the file on the drive letter of your choice.

After WinSPIRS is installed on the network drive, workstations must have that drive letter mapped. From the workstation, run the file, setup.exe located in the WinSPIRS directory. This sets up the workstation with the necessary program files.

To configure WinSPIRS 4.0 to retrieve the database information, you may choose to have the CDs available through a CD tower, or copy the data to the hard drive of a shared PC. In either case, you must identify the path needed to locate the databases, for example:

\\servername\volume\directory\databaseID

You need to configure the file, WINSPIRS.INI, located in the WinSPIRS directory, as follows:

1. Change to the WinSPIRS directory and edit the WINSPIRS.INI file using a DOS text editor.
2. Move to the section entitled [CORE] in the WINSPIRS.INI file.
3. Add the path to locate the area of the network where the data is located. The * wildcard may be used to choose all databases located in an area, as follows:

DRIVES=\\servername\volume\directory\*

4. Ensure that the [CORE] section of WINSPIRS.INI also contains the parameter:

MAGDRIVE=1

Save and close the WINSPIRS.INI file.

WinSPIRS is displaying an unreadable font (wingdings) on the screen or when printing records.
WinSPIRS uses the Courier–LD font as the default display and print font. If this font is not present in the system directory, unreadable characters will be displayed.

To resolve this problem:

1. If WinSPIRS is installed as a shared application, make sure the workstation setup program has been run. This is done by clicking on setup.exe in the WinSPIRS directory on the server.
2. Try using the WinSPIRS reconfiguration program to change the default print or display font. Within the Workstation Options and Print Options sections, there are buttons for Font and Fixed Pitch Font, which will allow you to designate a different font.
3. Copy the crld.ttf file from the WinSPIRS/shared directory to the windows\system or winnt\system directory and restart the workstation.
WinSPIRS should now display and print records legibly.

WinSPIRS can be configured to log into an ERL server without requiring the user to enter a username or password.
Directions for WinSPIRS v 4 and greater:

1. Edit the ERL Config file, located in your WinSPIRS directory (click start, programs, winspirs).
2. Make sure the WinSPIRS mode on the right had side is checked to local and ERL.
3. Highlight the example server name and click Remove Server.
4. Type in your servers name or IP address and click on Add Server. (example erl2.silverplatter.com).
5. Click Save.
6. Click on Run Reconfig.
7. Click on ERL Options.
8. Make sure WinSPIRS mode is again checked to local and ERL.
9. Under Automatic Login click Add.
10. Leave Server ID blank and fill in the username and password.
11. Click OK.
12. Click OK again at the top right hand corner.
13. Click on Save.
14. Click on Exit.

Directions for WinSPIRS v 2
1. Click on the start button using the right side of the mouse.
2. Choose Find.
3. Search for erledit.exe in the C: directory.
4. Double click on the file name erledit.exe.
5. Highlight the example server name and click Remove Server.
6. Type in your servers name or IP address and click on Add Server. (example erl2.silverplatter.com).
7. Make sure the WinSPIRS mode on the right had side is checked to local and ERL.
8. Click Save.
9. Click on Run Reconfig.
10. Click on ERL Options.
11. Make sure WinSPIRS mode is again checked to local and ERL.
12. Under Automatic Login click Add.
13. Leave Server ID blank and fill in the username and password.
14. Click OK.
15. Click OK again at the top right hand corner.
16. Click on Save.
17. Click on Exit

Is it possible for me to copy data from SilverPlatter CDs to my local hard drive and use WinSPIRS to retrieve the data?
Yes, instructions follow.

WinSPIRS can be configured to locate the databases on your local C: drive as follows:

1) Create directories for each database. – You may choose to create a C:\data directory off the root of C:. Then, for example, if the data you were installing was Poltox, you would create 7 directories underneath C:\data that would be:

P19A
P19B
P19C
P19D
P19E
P19F
P114

2) Copy the data from each corresponding CD to the directory you created.

3) Edit winspirs.ini located in the C:\winspirs directory, as follows (in the [CORE] section):

drives=C:\data\*

4) Now test WinSPIRS.

Configuring WinSPIRS for SDIs and Alerts
Note: WinSPIRS 4.0 uses WebSPIRS 4.0 functionality in order to utilize the SDIs and Alerts. Without access to a WebSPIRS gateway, you are unable to utilize these options.

1) Open ERL Config (by choosing Start, Programs, WinSPIRS, ERL Config).
2) Under "WinSPIRS mode", click "Local and ERL"
3) Under "ERL Servers", and "IP Address or Host Name", enter the name of the ERL server here – For example: erl2.silverplatter.com.
4) Press "Add Server"
5) Highlight (Under "ERL Servers") "your_server_name_or_ip_address_here/416" and press "Remove Server"
6) Press "Run Reconfig".
7) Choose "SDI/Alerts Option"
8) Under SDI/Alerts Permissions, check the boxes:

Allow SDI creation and Access
Allow Access to Alerts Service

10) Enter the URL of your WebSPIRS server under, "WebSPIRS URL for SDI/Alerts", for example:

http://webspirs.silverplatter.com:8000/cXXXX

11) Check the box "Login to WebSPIRS Automatically"
12) Enter your username and password for WebSPIRS
13) Press OK, then press SAVE at the main "Configuration Options" screen.
14) Press SAVE at the ERL Configuration Editor screen.
15) Now test – Access a database of your choice and run a search that you would like to set up as an SDI or Alert. For example, search the term "test". After pressing "Search", the records will be displayed. (In order to set up an SDI, you must first have completed a search.)
Now choose "File" and "Create SDI" from the menu (or "File", and "Alert Service"). Familiarize yourself with the screen and create the SDI.

WinSPIRS displays "Margins too wide error" when I attempt to print using Windows NT.
NT naming conventions can cause problems when printing from WinSPIRS. Printers with long UNC printer names are not understood by WinSPIRS. One solution is to set up a new printer icon for WinSPIRS:

Steps to set up a printer:
* Choose to add new printer
* Select 'My Computer' radio button, rather than 'Network Printer Server' button
* Under 'available ports', select the one with a UNC path or IP address path to your network printer.
* Answer the questions regarding your network printer – manufacturer and model.
* It will prompt you for the disk or path to load the printer driver. You may need to insert the NT disk or network path. (The program may already find it on your machine.)
* Choose 'not shared' and accept the options to complete the wizard.

You have selected an illegal drive. The valid drives are ... error message.
The WinSPIRS software can be configured to download to only specified, or legal, drives. Although the Download options allow the user to select any drive letter on the system, the actual downloading is only allowed on the drives specified in the Reconfiguration program.

To configure legal drive letters for downloading:

1. Double–click on the WinSPIRS Reconfiguration icon.
2. Click once on the FILE OPTIONS button.
3. In the Record Options section of the File Options dialog box, move to the ALLOWABLE DRIVES parameter. List those drives that may be used for downloading and separate them by commas. For example: A,B,C
4. Click once on the OK button to exit the File Options dialog box.
5. Finally, click once on the SAVE button to save the change.

'Show Held Titles Only' option incorrectly reports 'No titles found'
This problem is a bug in WinSPIRS v4.0 (9802).

The problem does NOT occur under the following conditions:– when one database segment is selected (for example, MEDLINE 1998)– when WinSPIRS v4.0 is used to search databases on an ERL server (v2.09.5, 2.11, 4.04) – regardless of the number of database segments selected

The problem DOES occur under the following conditions:– when more than one database segment is selected. This includes selecting mutliple years of the same database (for example, MEDLINE 1998 and 1997) and selecting multiple databases (for example, MEDLINE 1998 and CINAHL).
AND
– when WinSPIRS v4.0 is used in a standalone or local area network environment (non–ERL)

The result is a message that reports 'No titles found'.

Please note that the holdings messages do display correctly within the records so your users will see holdings messages for those records that contain them. It is the function of showing held titles only that is not working correctly.

This problem will be corrected in the next release of WinSPIRS.

I have WinSPIRS networked and I am getting "no databases found." What can I do to enable WinSPIRS to find the databases?
"No databases found" is a general error and does not indicate where, specifically, the problem lies. Below are some troubleshooting steps to follow:

GENERAL INFORMATION: WinSPIRS looks at the mapped drives of the workstation PC in order to find databases to load. It also looks at the DRIVES= line of the winspirs.ini file located in the WinSPIRS directory.

1) If you have just installed WinSPIRS, to verify that WinSPIRS is working correctly, place a SilverPlatter database in the local drive of the workstation PC (You may use the Software Resource CD to test. It has a small database, "SilverPlatter Electronic Directory" which can be used to test). Run WinSPIRS and verify that the program can find the database locally. If not, check the following:
a) Check other discs in the CD drive. Is the drive itself working correctly?
b) Locate the WinSPIRS directory on the networked drive. Run the program 'setup.exe' on the workstation. Now run WinSPIRS to verify that the program can find the local CD.

2) Once you have verified that WinSPIRS runs properly with a CD in the local drive of the workstation, verify that you are able to map a drive and run WinSPIRS from a mapped drive, as follows:
a) Open winspirs.ini, located in the WinSPIRS directory on the network drive. Scroll down to the line in the [CORE] section: DRIVES=
b) Remove any entries in this line, or remark the line.
c) Within Windows Explorer, (Tools, Map Network Drive) map a drive letter directly to the directory where one of the databases is located. Run WinSPIRS and verify that the program can find the database across the network.

3) Once you have verified that WinSPIRS runs properly with a mapped drive letter, you may choose to enter information regarding the location of multiple databases in the DRIVES= line of the winspirs.ini file.
a) Since you have successfully mapped a network drive, remove the mapped drive and make note of the exact path used when mapping.
b) In the WinSPIRS directory, on the network drive, edit the winspirs.ini file using the following format:

DRIVES=\\servername\volume\directory\subdirectory

To edit the line to go to one database (say, Ageline), example follows:

DRIVES=\\server1\sys\data\AL09

If you have multiple databases (say, Cinahl, Medline, and Ageline) all in directories within the 'data' directory, you could find all databases at once with the following line:

DRIVES=\\server1\sys\data\*

If you have databases located on two different servers, you could enter the information as follows with a semicolon separating:

DRIVES=\\server1\sys\data\*;\\server2\vol1\data\*

Please contact Technical Support if the above suggestions do not resolve your problem. If you identify at which step above (1, 2, or 3) the problem occurs, this information will assist in resolving your situation.

ininst caused a gpf in _ins042.mp_ when installing WinSPIRS on Windows98
This error is caused by a conflict between the Installshield program and a process which runs on Windows 98. To resolve the problem, press Ctrl–Alt–Del to bring up the list of processes running in the Task Manager. Look for either of the following processes:

mswheel.exe
EZKEY.exe

If either or both of these processes are running, click once to highlight the process, then click on the "End Task" button. This will disable the processes. At this point you will be able to run the WinSPIRS installation. Restarting the workstation will restore any processes that have been manually terminated.

Once installed, WinSPIRS will not conflict with either of these processes.

Assert Failed UTIL\ARRAY.C Line 1700
This error may occur when starting WinSPIRS when a portion of the winspirs.ini file becomes unreadable. Reinstalling WinSPIRS will not always resolve the problem because the WinSPIRS installer does not automatically overwrite the winspirs.ini file when WinSPIRS is reinstalled into the same directory. Either manually delete the winspirs.ini file before reinstallation, or install WinSPIRS into a directory other than where it was previously installed.

WinSPIRS will not install
If you have troubles installing WinSPIRS and you want to exclude all factors that could cause problems during install do the following:

– From you current WinSPIRS directory, back–up the files with the extension .his and other files that you want to keep. (only for WinSPIRS upgrades)
– Uninstall the current WinSPIRS installation
– Reboot your workstation
– Make sure that no programs (not even MS Office tool bar) are running when trying to install WinSPIRS. If the toolbar for Microsoft is running an error message will be given and the installation will not proceed, so close it down and try to install again, this time it should go ahead and you should have got rid off the error message.
– When the installation completes reboot the machine and start WinSPIRS.

I am not able to install WinSPIRS to my Windows 98 PC. I am getting the error, "The program has performed an illegal operation and will be shut down" when I try to install. What should I do?
In order to install WinSPIRS on Win98, press the Start button, choose Shutdown, choose 'Restart in MS Dos Mode'. When the computer restarts enter at the DOS prompt:

Rename c:\windows\system\mswheel.exe mswheel.old

Now restart the computer into Windows 98 and install WinSPIRS.
Go to the DOS prompt and type:

cd system
Rename mswheel.old mswheel.exe

I would like to install WinSPIRS using the SilverPlatter Resource CD. Which steps should I follow?
1) Choose "Software and Documentation" within the left–hand frame
2) Scroll down and click on "Software Directory" at the bottom.
3) Click on the "winspirs.40" folder.
4) Click on "win40.exe" – "Save it to disk" in a directory you choose, then run it to install into the WinSPIRS directory.

Note: If you are using Windows 98 and are receiving an error during installation, please click here for assistance.

How Do I download and install WinSPIRS?
WinSPIRS may be downloaded through our website.

If you choose to download WinSPIRS over the Internet, please click on the following link:

http://www.ovid.com/site/support/sp_software_download.jsp
Place an "X" in the box in front of 'WinSPIRS'.

Complete the registration information. (It is only necessary to complete this information once. It will be carried over in the future.) Agree to the licensing information.

You may choose to download the documentation (WinSPIRS (User's Manual) – pdf. This requires Adobe Acrobat in order to read the information.)

The installation file is WinSPIRS for Windows in English.

Double click here. This will download into a directory specified by your computer. Make note of the location of this download.

Locate the win.exe (filename will be win(version number).exe for example win401.exe) file using Windows explorer.

Double click on the file to run it.

This will start an Install Wizard.

You will be prompted with various questions during install.

If you are confused, or do not know the correct answer, please email Technical Support at support@silverplatter.com. The program will install and appear in the "Programs" area of your computer (Win95/98).

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

MACSPIRS

General


Where do I download a Year 2000 comliant version of MacSPIRS?
MacSPIRS v4.05 is Y2K compliant and is available for download from SilverPlatter's anonymous ftp site. To download the software:
1. Use anonymous ftp to connect to SilverPlatter's site: ftp.silverplatter.com

2. Change to the /software/macspirs4.05 directory

3. Set the transfer mode to binary by typing: bin

4. Download all files from the directory by typing: mget * (you will be prompted to entery y to agree to download each file)

The files included in this directory are:

MacSPIRS_40k.sea.hqx – MacSPIRS software
macman40.pdf – MacSPIRS Reference Manual (PDF file)
qrcmac40.pdf – MacSPIRS Quick Reference Card (PDF file)
relnotes.txt – Release Notes (text file)

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

Z39.50

General

Error Messages


Directions to create a connection file with EndNote.
You must have a Z3950 Gateway running in order to access SilverPlatter databases using the EndNote client. The Z3950 Gateway must be running and configured for a port (default is 210) and the ERL server you are trying to access.

Open EndNote and from the main screen choose "File", "Connection Files", "New Connection File."

Enter the necessary information as follows: Server Description: YOUR SERVER (Server where Z39.50 Gateway is located)

Server Address: YOUR SERVER (Server where Z39.50 Gateway is located)

Database Description: MEDS008A (Or the exact name of a database located on the ERL server)

Database Name (examples): MEDS008A or MEDS* or PSYI* etc.

Element Set Name: F Port ID: 210

Record Syntax: USMARC

[ X] Password:

[ X ] User ID:

Note: Check off the Password and User ID boxes, but do not enter the information in the boxes if you'd like the user to be prompted to enter the username and password upon attempting to connect. If you'd like to automatically login, enter the username and password that you use to connect to the ERL server.

Press the "X" in the upper right corner to close and you will be prompted whether you wish to save the file or not. Choose to save the file and choose a name for the file. Now that the file has been saved, you may connect using this configured information by choosing "File" and "Connect" and double–clicking on the filename you created.

Where do I find information on how different SilverPlatter databases have been mapped to MARC format?
SilverPlatter's Z39.50 Gateway provides precision searching and full database content in the context of standard Bib–1 attributes and USMARC records. Simple searches and basic record display can be done in a completely database–independent way using basic attributes like Title, Author, ISSN, etc. and generic MARC field names. Results comparable with SilverPlatter's interfaces can only be achieved with database specific configuration. Most of the relevant information is available through the Gateway's Explain database. But since very few clients support Explain, we provide these pages to support manual client configuration.

MARC Map file (MFF) information and download section

In my Z39.50 Client, I am receiving the following Error:
Error_code=227, message='No data available in requested record syntax'recordSyntax=''

This error indicates that the necessary .mff file is not present. In order to correct this problem, complete the following:

1) Go to the Download section

2) Click on the database link whose .mff file is needed and download the corresponding .mff file.

3) Copy this file to the /z3950/recmap directory on your server.

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

SILVERLINKER

Overview

General

Configuration Options


Font problems when printing Silverlinker records
Displaying full content providers in a drop–down format within records can cause full content Silverlinker links to print in a small font. To eliminate theproblem, change the configurable setting by editing the webspirs_user.cfg file and changing the following parameter:

format=SELECT
to
format=TEXT

This will cause records with multiple full content providers to display the providers in a list format rather than a drop–down box, and should eliminate the font problems when the records are printed.

What are the current SilverLinker Publisher URLs?
As of March 2000 they are:

Publisher url_prefixes

Academic Press http://www.idealibrary.com/
Academic Press http://www.europe.idealibrary.com/
CatchWord http://www.catchword.com/
DeGruyter http://www.degruyter.com/
Ebsco http://www.ebsco.com/
IEE http://ioj.iee.org.uk/
Information Quest http://www.eiq.com/
Karger http://www.online.karger.com/
Kluwer Academic Publishers http://kapis.www.wkap.nl/
NRC Research Press http://www.nrc.ca/cisti/
Project Muse http://muse.jhu.edu/
Science Direct http://www.sciencedirect.com/
SIAM http://epubs.siam.org/
Springer http://link.springer.de/
Springer http://link.springer–ny.com/
SwetsNet http://www.swetsnet.nl/
SwetsNet http://www.swetsnet.com/

Configuring SilverLinker Fulltext Prompts in ERL 4.06
When users search a SilverLinker–enabled database with WebSPIRS 4.1 or higher, WinSPIRS 4.01, or MacSPIRS 4.05 links to full text articles offer prompt text identifying the journal publisher. For publishers added to SilverLinker after the release of ERL 4.06 it is necessary to add the prompt text by editing erlsrvr.cfg.

To configure the prompt text, edit the entries in the [pmt] section of erlsrvr.cfg file located in the sproot/bin directory.

A prompt entry has the form:
URL_hostname = prompt_string

Configuring SilverLinker URL_hostname is the base address of the Web site, including the protocol prefix and trailing slash. prompt_string is any single line of text:

this is the text displayed on the screen.

Note:
Enter each line as a single line of text.

For example:
[pmt]
http://www.journalpub.com/ = Full Text from Journal Publishers
http://www.europe.journalpub.com/ = Full Text from Journal Publishers (European Mirror)

This example displays "Full Text from Journal Publishers" or "Full Text from Journal Publishers (European Mirror)" in the FTXT field.

For a list of current SilverLinker publisher URLs please click here

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

WEBADMIN

Statistics

General

Configuration Options

Installation


On WebAdmin a user id with authorization to see stats gets no data shown when a report is run
First make sure that you have accessed few databases with that user id, secondly you can run the stats report while logged on WebAdmin with the user id that has the administrative right for the stats, make sure it includes the last day that a database was accessed with that user id. Also, note that anyone using WebAdmin with an ERL User Account that has been authorized to see statistics can only view statistics for that account and cannot choose to see Statistics for another Account.

If user id needs to see and run the statistics for all users that access the databases, then the user id needs to be given Guest Administrator rights. This will allow that user id to see everything (as if it was a Full Administrator) but can only run the statistics option in WebAdmin.

Could you give me a summary of the steps to follow to create a statistics report using ERL v4 and WebAdmin?
WebAdmin allows you to generate a wide range of statistical reports against your ERL server. You can focus your reports on accounts, databases, portfolios, or IP addresses, or any combination.

1. Log in to ERL with admin/admin, or other administrative account.
2. Click Statistics. WebAdmin displays the Statistics window, which lists all the statistical reports you have saved with the current account.
3. Click New. WebAdmin displays the New Report window.
4. Enter a Name and Description for the report.
5. Select the date range for which you want to collect statistics.
6. Select the Style, or granularity of the report. The options are Per Session, Hourly, Daily, Monthly, and Totaled.
7. Specify whether you want the statistics sorted chronologically or alphanumerically.
8. Select zero or more Key Fields to feature in your report. The possible key fields are Account Name, Database ID, IP Address, DB Group, and Portfolio. (If you select no keys, you generate information about all activities.)
9. Select the Data Items you want to include in the report.
10. Click Run Now. WebAdmin retrieves statistics from ERL and displays the Statistics window.
You can then print the report, or export it to use in a spreadsheet. Refer to the WebAdmin online help for more detailed information about statistics.

I have installed WebAdmin for Windows NT. Under IIS version 3 or 4, what should the virtual directories be set to?
These are the default settings after completing the Installation of WebAdmin for Windows NT.

Open Internet Services Manager, these directories should be set up as virtual directories

a. C:\InetPub

ALIAS=/CInetPub

b. C:\Program Files\SilverPlatter Information\WebAdmin\scripts

ALIAS=/wascripts

read and execute attributes: should be checked off

c. C:\Program Files\SilverPlatter Information\WebAdmin

ALIAS=/webadmin

read attribute: should be checked

Please provide some easy–to–follow instructions regarding BRLAdmin installation and how to run the Solaris script.
NOTE: BrlAdmin is used to run against stat files that are currently running on the ERL server. If you are trying to use this for old stat files that are large or possibly corrupt, you must shut down your ERL service (choose a time when no users are logging in), place the old stat files up on the server, then restart the ERL process. The files are read through the ERL server. If you do not wish to place these up on the server, your alternative would be to use LISTER or READER.

Solaris Instructions:

1)Login as root and download the package from ftp.silverplatter.com. It is located in the /software/erl–administration/brladmin directory. Choose the platform. Download as binary. Use the mget * to download the three files:

brladmin.pkg
readme.txt
test.bsx

2) Install the package using the following:

pkgadd –d /exactpath/brladmin.pkg

3) You will be asked, during install, where to place the files. I placed them in /sproot.
4) If you have installed successfully, a brladmin directory and brlscripts directory in /sproot will now display. You will also have an executable brl and brladmin.cfg in /sproot/bin.
5) Edit the test.bsx file, located with your initial download area. I edited the username and password to be admin and admin and I also needed to edit the path where the file to be created will be located. (The sample test.bsx contains information as if it will send the information to a D: drive.) You may edit the various headings you would like included in the resulting stat report.
6) Copy the newly edited test.bsx to the same directory as your brl executable. (In my case, I copied it to the /sproot/bin directory).
7) Run the script as follows:

brl –r999999 test.bsx

(The 999999 represents time between retries.)

8) This will now create a new txt file which has been run against your current stat files on your server. The file is named stats.txt. (Mine was located in my /sproot directory.)

I have installed WebAdmin, but when I try to access it my server is unable to find the software.
In the /sproot/apache/conf/httpd.conf file, add the following entry:

ServerName IP

Where you replace "IP" with the IP address of the ERL server (or servername).

On the Solaris platform, the following (additional) information should be added:

Port 4416
User erl
Group erl
ServerName {IP or DNS}

Additional questions or problems should be directed to Technical Support.

Admin install error
Error: The Web server cannot find the file or script you asked for.

When trying to install WebAdmin onto an NT Server, if the following error message displays:

ERROR:HTTP Error 404
404 Not Found
The Web server cannot find the file or script you asked for. Please check the URL to ensure that the path is correct.
Please contact the server's administrator if this problem persists.

The problem lies in the fact that a Perl5003 directory must be present in the WebAdmin directory. If a previous copy of WebAdmin was present on the machine, and the Registry thinks that Perl5003 is already present, WebAdmin will install without the Perl5003 directory. To correct this situation, you may copy the Perl5003 directory from the previous installation. Otherwise, delete references to Perl5003 from the Registry, using Regedit, remove the present installation of WebAdmin, and reinstall.

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

HCU

Overview

General

Configuration Options

Error Messages

Installation


Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.

Placement of Holdings files on ERL v4.x server
HCU (Holdings Conversion Utility) creates three holdings files that must be placed on the ERL server in order for software clients to read library holdings messages. These three files are:

SPLHA.LHD
SPLHA.LHI
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

SPLHA (SilverPlatter Information Library Holdings Administrator) creates two holdings files that must be placed on the ERL server. These two files are:

SPLHA.LHD
SPLHA.LHM

These files should be placed in the sproot/cfiles directory of the ERL server. If the SPLHA.* files were not created on the same machine that hosts ERL, ftp the files in binary mode to the ERL server and place in the sproot/cfiles directory.

It is also possible to have different holdings files for specific ERL users by placing the SPLHA.* files under user specific subdirectories of sproot/cfiles, i.e., sproot/cfiles/00000004/. The user specific subdirectories are named after the user Ident found in ErlAdmin – an 8–digit number. In WebAdmin the preceding zeroes are assumed – use the ID number and preface it with enough zeroes to comprise 8 digits.

Which holdings option is better – HCU/static file solution or WebLink to Holdings/dynamic solution?
Each holdings option has advantages, especially depending upon the environment in which you provide access to your SilverPlatter databases.

Listed below are some facts about each, which should help you decide which one makes most sense for your situation. NOTE that the two options can also be used together in some situations. For a detailed discussion of each holdings option we recommend that you read the SilverPlatter Holdings Documentation available at ftp://ftp.silverplatter.com/software/hcu.

HCU/Static File
– works with WebSPIRS, WinSPIRS and MacSPIRS
– message is displayed right along with SP record so users see needed information immediately
– as long as you can provide an input file of your serials holdings in the appropriate format, you can make holdings information available with HCU
– holdings information is included in any output of SilverPlatter records in which a message is displayed
– requires that you be able to produce a file of your serials in MARC or ASCII format, and update it as necessary
– works only for SilverPlatter serials–based records that contain ISSN numbers
– WinSPIRS "Limit to Holdings" feature requires the static holdings file (future versions of WebSPIRS 4 will also be able to limit results based on the static holdings file)

WebLink to Holdings/Dynamic
– works only with WebSPIRS v.4x
– requires Web–compatible OPAC with specific type of URL–protocol
– allows user to see most up–to–date and detailed information
– user needs to leave WebSPIRS record to see holdings
– requires library to have cataloged serials records in their OPAC
– provides access to holdings information for both books and serials based on ISSN or ISBN number contained in the SilverPlatter bibliographic record
– there is no easy way to print the holdings information as displayed from the OPAC

What would I use as the input file for HCU? (SilverPlatter Holdings Documentation Section 1.3)
Libraries which have cataloged their serials for inclusion in their OPAC's may be able to extract or export a file of these serials records from the OPAC in either MARC or ASCII format. If you're not sure about how to do this, you should talk to your OPAC vendor about the tools which are available for this type of reporting or extraction. Some libraries may be able to obtain this information from their serials subscription agent. Libraries which participate in a Serials Union List project may be able to utilize the information prepared as input for the Union List.

If you cannot extract this information from your OPAC or obtain it from your serials subscription agent or a Union Listing project, then the simplest way to create a serials holdings file for use with HCU is to use any standard word–processing, database; or spreadsheet software package to create an ASCII file which conforms to the specifications described in this document and which contains information about each of the serials titles in your collection.

You might also be able to obtain a file of your serials holdings from your subscription agent or from a serials union list project which your library participates in. The file must conform strictly to the requirements described in the documentation.

I have my library's serials data in Excel (or Word or WordPerfect, etc) format. How can I convert it to a delimited ASCII format to use with HCU. (SilverPlatter Holdings Documentation Section 1.3)
Use the application's functionality to save files as ASCII text. For example, in Microsoft Word, use the "SaveAs" facility, selecting the .txt format, to convert the file to an ASCII text file. In Excel, select Text (Tab Delimited)*.txt.

We've followed the instructions for including a SilverPlatter–style hotlink in our HCU–generated holdings files, but have not been successful in launching a browser to connect to the link (to our union list of serials).
It's essential that you follow the format for specifying the link information exactly as described in the HCU documentation. http://www.ovid.com/site/support/system_admin_manuals.jsp

AOL USERS ARE EXPERIENCING PROBLEMS ACCESSING WEBSPIRS
If you are using AOL to access SilverPlatter databases, you may experience problems. The error message that others have reported is "invalid user".

Although AOL appears to be using Internet Explorer, or Netscape, the browser is actually proprietary. This browser has not been tested in conjunction with our database search software. You need to use either Netscape or Internet Explorer outside of your AOL window, once you have connected, to access.

For more detailed information regarding browser requirements, please note the information under "Workstation Requirements" at the following link:

http://www.ovid.com/site/support/erlgatewayreq.jsp

How do I convert my present SPLHA library holding files so that I can use them with the new HCU program?
Once you have downloaded the HCU program, reverse.pl is a file located in the HCU directory. It is this program that enables one to convert SPLHA files into a .txt format, to be used with HCU. Instructions follow:

REVERSE.PL INSTRUCTIONS

1. Place the splha.lhd and splha.lhm files in the same directory as the reverse.pl file, which is, by default, the C:\Program Files\SilverPlatter Information\HCU directory. (You do not necessarily have to place the files there, but if you do not, you would identify the path of the files when you run reverse.pl below. As an example, the following information is included in the files:

Jama – This title is held in paper and microform Call Number: Testing

2. Run reverse.pl with the following command (Note, HCU is loaded on a Windows 95 workstation in this example. Please refer to the HCU documentation for running reverse.pl on other platforms.):

PERL5003\bin\perl –I PERL5003\lib reverse.pl splha.lhd splha.lhm –o outputfilename –l logfilename

3. View the log file and note the entries. The errors noted in the log file DO NOT need to be adjusted in the outputfile.

4. Enter the necessary information and run HCU using the "Convert ASCII to SPHoldings" Tab. You will received a set of newly created files:

splha.dat
splha.lhd
splha.lhi
splha.lhm
splha.log

Backup these files.

5. Place the splha.lhm and splha.lhd in the WinSPIRS directory, if you are using WinSPIRS. If you are using ERL, place these in the \sproot\cfiles directory.

6. To configure WinSPIRS for library holdings, start the Reconfig program. Click on the button, "Library Holding Options." Check the box, "Holdings Exist". Identify the path to your WinSPIRS directory where you have placed the splha.lhm and splha.lhd files.

7. Test. Open the WinSPIRS program (or in your browser, connect using WebSPIRS) and enter a search to display records where you know holdings exist. In the above case, enter the search:

JAMA in so

The following appeared in the record area of the example above:

TI: JAMA patient page. Weighing difficult choices: treatment of prostate cancer
SO: JAMA:–Journal–of–the–American–Medical–Association (JAMA) 1998 Sep 16; 280(11): 1030
*LHM: CNTesting This title is held in paper and microform.
SI: J03340000
AN: 1998062146

The "CNTesting" refers to the call number information

I used HCU to convert my serials file, and in general, my users see the appropriate holdings message for the cited journal. However, there are occasions when a SilverPlatter record which cites a journal which my library subscribes to and for which I have created a holdings message displays no message. Why might this be? (SilverPlatter Holdings Documentation Section 1.7)
There are a number of possible reasons for this. It is possible that the ISSN in the SilverPlatter record does not match the ISSN for the cited serial as it appears in your holdings record. This could be due to differing editions of the same journal; a change in the ISSN number; the fact that the SP record doesn't contain an ISSN number; or data entry error either in the SilverPlatter record or in your record.

In some cases, these discrepancies can be addressed through making additional entries in your ASCII file, or, if you run HCU on the UNIX platform, through use of the fixup file. In some cases, title matching may compensate for ISSN mismatches.

What is the maximum length for holdings message which can be converted by HCU and displayed with SilverPlatter databases?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), there is no limit on the length on any individual holdings message. With an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

How many messages can be supported by HCU?
When used with an ERL v.4 server; or with a v.4 standalone system (MacSPIRS v.4 or WinSPIRS v.4), a total of 64000 individual records can be supported. Each of these records may have a unique message if desired. There is no limit on the length on any individual holdings message. However, with an ERL v.2x server or standalone system, the entire resultant holdings file (.LHM) may not exceed a total of 64K bytes (i.e. 64,000 characters).

I would like a single message to display in the record for all SilverPlatter records, to refer our users to the printed list of serials which we carefully maintain. Do I need to enter this message individually for each of my journals? (SilverPlatter Holdings Documentation Section 1.4.1)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish.

I currently use the SPLHA (SilverPlatter Library Holdings Administration utility to display holdings for my SilverPlatter databases. Do I need to switch to HCU? (SilverPlatter Holdings Documentation Section 1.8)
There are many important reasons why libraries should plan to migrate to HCU as soon as possible. Here are just a few. SPLHA only supports a very limited number of SilverPlatter databases while HCU supports all ISSN–based SilverPlatter databases, any customers subscribing to databases in addition to those supported by SPLHA will want to begin using HCU in order to make holdings information available as widely as possible. HCU allows much more flexibility in terms of the number of different messages which can be supported. HCU–generated files only need to be updated when your library's serials collection has changed, while SPLHA requires updating every time there's a new database update. Finally, SilverPlatter will only continue to provide support for the SPLHA utility through 12/99. Following that date, there will be no way to update your SPLHA files to reflect changes in journal indexing within the SPLHA–supported databases.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. All of the libraries share a single ERL server housed at one of the institutions. We would like our users to see only the holdings of their own library displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

I'm a member of a 5–library consortium which would like to utilize HCU to create holdings files. We would like our users to see the merged holdings of all of the libraries displayed within the SilverPlatter record. Is there a way to accomplish this?
This can be accomplished as long as the users associated with each institution have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory. The result would be that users at library1 would see only holdings of library1; users at library2 would see only holdings of library2, and so on.

My library accesses SilverPlatter databases which are stored at 3 different locations, e.g., a local ERL server, a shared ERL server maintained at a different library and SilverPlatter's Norwood–based internet service. What do I need to do to have holdings display with any SilverPlatter database, regardless of the server on which the database is installed?
You would need to have your HCU–generated holdings file installed on each of the servers which your users access for SilverPlatter databases. In the case of the database accessed from a different library's server, your users would need to be using an ERL username/login which identifies them with your institution and distinguishes them from users of the other library.

My holdings messages do not always display in the same part of the record in my various SilverPlatter databases. Why?
With the exception of WebSPIRS, all of the SPIRS clients will display the LHM field containing the HCU–generated holdings message, following the SO or JN fields. WebSPIRS, however, displays the LHM field at the bottom of the brief or full record display, to facilitate the complementary use of the HCU holdings and WebLink to Holdings features.

I have a collection of full text articles loaded onto a website which my users can access. Is there any way to provide a link from a SilverPlatter record to the corresponding article in this collection?
If you have installed the newest version of the ERL server, v.4.0, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using a v.4 client (WinSPIRS, WebSPIRS, or MacSPIRS when available). The links could be made to a specific journal but not to a specific article within that journal.

Can I use the new Holdings features to provide holdings information about books in my library's collection?
Only the WebLink to Holdings supports holdings information for books, based on the ISBN number in the SilverPlatter records. HCU–generated holdings files can display only journal–related information, based on an ISSN. The limited size of the holdings files which can currently be accommodated on the ERL server precludes the inclusion of the generally much larger number of records which represent the library's book collection.

I've heard that my library has to have a Web–compatible OPAC in order to take advantage of SilverPlatter's holdings solution. Is this true?
ERL Release 4 provides two different ways to display local holdings information. The dynamic WebLink to Holdings requires a web–compatible OPAC and WebSPIRS v.4.X. However there is also the ability to display holdings information from a static file of serials information which has been converted using the Holdings Conversion Utility (HCU). This may be used with WebSPIRS, WinSPIRS or MacSPIRS.

How can I tell if my web–compatible OPAC will work with the WebLink to Holdings?
In general, the OPAC must conform to the following guidelines in order for the WebLink to be successful:

1. The Web OPAC allows direct searching in the catalogue on ISSN/ISBN number. i.e., the ISSN/ISBN numbers must be indexed in the catalogue database.

2. The URL that is used to search in the catalogue is pre–definable in such a way that only the ISSN/ISBN number has to change.

As a consequence of the above, the WebLink is unlikely to be compatible with Web OPACs that use dynamic "session–id's", dynamic "user–id's", or dynamic "port–numbers" in the URL. (Dynamic means that the session–id, user–id, or port–number in the URL differs in time. i.e. per browser session, per login etc.)

Note that the link will only work from SilverPlatter records that contain a field for ISSN or ISBN numbers. When creating the HTTP link, WebSPIRS will normalize the ISSN and ISBN numbers as follows:
– The ISSN number used in the URL will include the hyphen; example: 0021–9150
– The ISBN number will be normalized in the URL by dropping the hyphens, so 1–85898–469–6 will be normalized to 1858984696.

This may affect compatibility with WebLinking, depending upon how these numbers are indexed in the OPAC.

Which OPAC's are compatible with the WebLink to Holdings?
Customers have successfully implemented the WebLink to Holdings with the following OPAC's:

Ameritech Horizon/WebPac
BibSys
Endeavor/Voyager
Ex Libris/Aleph 500
Innovative Interfaces/Innopac
Libris
MiniSys
OCLC Web–Z/SiteSearch
PICA
Sirsi/Unicorn v. 98.4
Talis
TinLib

Coming soon...
VTLS/Virtua

NOTE though that different versions of the same vendor OPAC may not work the same way with the WebLink to Holdings. Also, it's possible that in some cases, libraries did some customization to WebSPIRS in order to enable the WebLink feature.

We've enabled the WebLink to Holdings option, but some SilverPlatter records don't display the Check for Holdings button. Why would this be?
Before displaying the Check Holdings button in a record, WebSPIRS first checks the record to see if there is an ISSN or ISBN number field by looking for one of the following short field names: IS, ISSN, ISBN, IB, or SN. If none of those fields are present in the record, then WebSPIRS will not display the button. Note that you can change the set of default fields which WebSPIRS checks as well as the priority order in which they're checked.

We'd like to implement the WebLink to Holdings but are concerned that our users will be wasting a lot of time linking to the OPAC only to find out that the cited material isn't owned by the library. Is there any way to avoid this?
If you implement both the HCU and WebLink to Holdings features, then the resultant holdings file can be used to serve as a filter which will cause WebLinks only to display in records for which your library has indicated holdings in the HCU file.

NOTE this filter is only available for journals, not for books since the HCU file only accommodates ISSNs while the WebLink to Holdings accommodates ISSNs and ISBNs. If the HCU files are used as a filter any records containing a separate ISBN field will not display the Check for Holdings button because a match will not be present in the holdings files.

To set up the HCU files to serve as a filter, you must set the associated parameter in the webspirs_user.cfg file SP.HOLDINGS section to 1 as shown below:

requires.text.p=1

Can I have SPLHA and HCU created holdings files on my ERL v4.x server?
It is not possible for both HCU created SPLHA files and SPLHA created SPLHA files to reside directly under sproot/cfiles nor a combination of SPLHA created files under sproot/cfiles and HCU created files under user subdirectories on the same ERL server. A connection lost message will occur when a user attempts to connect to the server if a combination of files is present.

If different types of SPLHA files are to be used they must all reside under user specific subdirectories.

Excluding Holdings Messages for Specific Databases – ERL v4.x
The default is for HCU generated holdings messages to display for all SilverPlatter databases that contain ISSN data in a separate field.

It is possible to exclude the display of HCU generated holdings messages for specific database families by creating a list of those database families within /sproot/dbase/coretags.cfg, line !42 and removing the semi–colon (;) preceding !42 as shown in the example below. This example excludes Popline and GPO from displaying holdings messages.

!42=PO,GP

Configuring HCU Default Messages – ERL v4.x
The coretags.cfg file also allows for the display of default messages in cases where a journal specific holdings message is not available.

The /sproot/dbase/coretags.cfg file contains the following list of numeric tags and textual messages associated with various conditions that result in no journal–specific messages available for display:

;!50=Holdings files missing or incomplete
;!51=This database contains no valid holdings keys
;!52=Holdings for this database family disabled
;!53=No holdings key field in record
;!54=No holdings key found in field
;!55=The holdings key in this record is invalid
;!56=Holdings index file corrupt – rebuild
;!57=No holdings information for this record

To enable a default message to display open the coretags.cfg file in a text editor.

Delete the semi–colon (;) at the beginning of the tag which corresponds to the condition for which a default message should display. By removing the semi–colon the message is activated. The text may be left as is or edited for a new or additional message. If messages span multiple lines, type \n at the end of each separate line of text to create line breaks; do not press [Enter]. If [Enter] is pressed at the end of a line, the message will be truncated at that point. Save the edited coretags.cfg file.

For example if a library does not subscribe to the journal that is cited in the SilverPlatter bibliographic record displayed, no HCU generated message would display. To activate a message to display for this condition use message !57.

I can extract a file of my serials holdings which contains the ISSN and title of all my journals, however, there is no way for me to extract a textual holdings message for the journals. Is there some way I can use HCU? (SilverPlatter Holdings Documentation Section 1.4; Section 2)
The HCU utility allows you to create default messages which will be displayed in the absence of a journal–specific message for a particular ISSN. You may use default messages for all of your titles, if you wish. Of course, the default message in this case would need to be a generic one which could apply to all of your journals. For example, you might create a default message which said "Library subscribes to this journal. Please check printed Periodicals List for details"

If you are in an environment which supports the WebLink to Holdings capability, you may also use an ISSN/title file created with HCU and installed on your ERL as a filter for the display of the WebLink to Holdings button only in those SilverPlatter records for which there is an ISSN match on your static file. This also involves setting the relevant webspirs_user.cfg options.

I maintain a table containing my serials holdings information on the web. Can I provide a link from SilverPlatter records to this website? (SilverPlatter Holdings Documentation Section 1.4.2; 1.4.1 )
If you are using the latest version of the ERL server, v.4, you can embed a URL in your holdings message, using the SilverPlatter–specific format described in the Holdings documentation. HCU will convert that information into a hyperlink; WinSPIRS v.4, WebSPIRS v.4 (and MacSPIRS v.4 when available) will display the information as an active link. Depending upon the structure of your website, you may be able to provide journal–specific links, although in many cases you may only be able to support a link to the top level of the website. In the latter case, you could use HCU's default message capability to provide a generic link to the website

My library has a web–compatible OPAC, and we'd like to implement the WebLink to Holdings. I've set up my configuration options as instructed, and the Holdings button displays in my records, but we cannot successfully connect to the OPAC. Why might this be?
The protocol which your OPAC uses for constructing a URL must conform to certain requirements. In the current implementation of WebSPIRS 4, the ISSN or ISBN number must appear at the end of the URL. Also, the URL must be static; i.e. if the OPAC appends a dynamic session ID to the URL constructed for each search, WebSPIRS will not be able to link to records in your OPAC. Also, the matching will not be very precise if your OPAC does not support explicit ISSN or ISBN searching. E.g. if the search query protocol supports only keyword searching for either of these entities, then the resultant match may be on any type of numeric entries in your OPAC records.

Section 2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

I have some workstations running WinSPIRS and some using WebSPIRS to connect to our ERL server. Which holdings solution is best for me?
In order to provide holdings information from any of your SilverPlatter workstations, you need to use the HCU utility to create static files of your holdings information, and place these files (SPLHA.LHM, SPLHA.LHI, SPLHA.LHD) in the sproot/cfiles directory on your ERL server. Holdings messages will then be available for display with both WinSPIRS and WebSPIRS. You may also enable the WebLink to Holdings option, but the Check Holdings buttons associated with WebLinking will only display when WebSPIRS 4.X is the retrieval client used.

I have some workstations running WinSPIRS and some running WebSPIRS. Is there any way to provide links to my OPAC from all of these workstations?
You could implement the WebLink to Holdings for searching SilverPlatter databases from WebSPIRS, as long as your OPAC conforms to the required protocols. In order to enable linking to your OPAC from WinSPIRS, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the Holdings documentation. (Section 1.4.2, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using WinSPIRS v. 4.X and ERL server, v.4.X.

Note that if you embed links in your HCU–generated messages, these links will also be available within WebSPIRS unless you disable the display of static library holdings messages (show.text.p=0) in the SP.HOLDINGS section of the webspirs_user.cfg.

I'm a member of a 5–library consortium. We would like our users to see both the holdings of their own home library as well as the merged holdings of all of the in the consortium. Is there a way to accomplish this?
If you use WebSPIRS 4 to access your SilverPlatter databases and if the consortium shares an OPAC which conforms to the specifications for WebLink to Holdings, you could use library–specific holdings displayed from an HCU–generated serials holdings file for each library along with the WebLink to Holdings, which would allow users to link to the OPAC record which contains holdings information for all libraries in the consortium.

The users associated with each institution must have ERL usernames/logins which associate them with their institution and which distinguish them from users at each of the other institutions. Each library needs to generate a file of its own serials holdings and convert those holdings using HCU. All of the resultant files would then be installed on the consortium's ERL server in the appropriate sproot/cfiles/user directory.

On the WebSPIRS side the SP.HOLDINGS section of the webspirs_user.cfg file should be configured to point to the shared OPAC. The result would be that users at library1 would see only static holdings of library1; users at library2 would see only holdings of library2, but WebLink to Holdings would go to the same OPAC for all users.

What SPIRS clients can use WebLink to Holdings and HCU files?
Only WebSPIRS v.4.X supports the WebLink to Holdings.

HCU–generated holdings messages can be displayed in WinSPIRS v.4.X or WinSPIRS v.2.1/2.2; WebSPIRS v.4.X or WebSPIRS v.3.1; and MacSPIRS v. 4.X or MacSPIRS v. 2.4. Certain features such as holdings messages containing URLs and certain databases such as ERIC require v. 4.X retrieval software and v.4.X ERL if used in an ERL environment.

Section 1.6, HCU Manual, http://www.ovid.com/site/support/system_admin_manuals.jsp

What databases are supported by the HCU and WebLink to Holdings features?
The HCU holdings feature provides a holdings message for any SilverPlatter database that includes ISSNs in a separate field.

Note, however, that many ISSN–based databases also contain some subset of records that do not include an ISSN number. For example, older segments of databases which include extensive backfiles are less likely to include ISSN's than newer segments. There are many databases that index a combination of serials and other types of material, and only the ISSN–based records can display a holdings message.

There are also databases that include a rubric (i.e. extraneous textual data) preceding the actual ISSN, or are produced with different technology. Records in these databases will not successfully display holdings messages unless used with ERL server v. 4.X or a v.4.X standalone system.

The WebLink to Holdings allows links to the OPAC in any SilverPlatter record which contains either an ISSN or an ISBN in a separate field.

I have noticed that different SilverPlatter databases index different editions of the same journal (e.g. US and British edition), which often have different ISSN's. Consequently, my holdings message doesn't always display. Is there anything I can do?
If you are using holdings files created with HCU there are several options:

1. Holdings ASCII file – add a new record showing the journal with its alternate ISSN to the ASCII holdings files, and convert again with HCU.

2. Holdings MARC records – make an entry in the Fixup File (only available for HCU/Unix), and convert again with HCU.

3. Holdings MARC records – make an entry in the ASCII file created by the MARC to ASCII conversion in HCU. Convert ASCII to SPLHA holdings with HCU.

Post these new splha.lhi, splha.lhm, and splha.lhd to the ERL server or standalone client.

For WebLink to Holdings work, you need to have each edition of the journal cataloged separately.

How often will I need to run HCU? How up–to–date will my holdings information be? Do I need to run HCU every time I get a database update?
You should run HCU as frequently as you feel is necessary to ensure that your holdings information is accurate. This will depend upon the frequency with which you add new serials titles or weed your serials collection. You don't need to run HCU when you get database updates, as the necessary information for HCU is based on your own serials holdings information, not on the serials in the SilverPlatter databases to which you subscribe.

The WebLink to Holdings button is linking to inappropriate records. Why might this be?
Because the field which contains an ISSN or ISBN in a SilverPlatter database does not always have the same field name, WebSPIRS is programmed to check in a small number of specific fields for the relevant data. This can occasionally cause conflicts. For example, the ISSN or ISBN number is sometimesin a field labeled SN, so this is one of the fields in which WebSPIRS checks. However, in a few databases the SN field contains an unrelated number such as a "sequence number". In some cases, for instance, if a particular SilverPlatter record contains no ISSN or ISBN number but the record contains an SN field with some other type of number in it, WebSPIRS may try to find a match in the OPAC based on this SN number, which would yield inappropriate results.If the WebLink button is not appearing as expected in one or all of the records in any of your SilverPlatter databases, you should check:(1) to see if there is an ISSN or ISBN number in the record(s) in question(2) if there is an ISSN or ISBN number, note the short field label of the fieldin which the number appears(3) compare that field to the list of fields in the "link.fieldlist.p="entry in the SP.HOLDINGS section of webspirs_user.cfg.Conflicts may occur depending upon the combination of SilverPlatter databases to which you subscribe and the fields in which the ISSN or ISBN appears in these databases. The following list illustrates some examples of databases in which the ISSN or ISBN numbers appear in a way which might lead to a conflict, and the possible actions which you might take.NOTE: this is not an exhaustive list, but is intended to be representative.A MATTER OF FACT (AMOF): ISSN is in IS field and ISBN is in IB field. However, AMOF also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field.RESOLUTION: If you have no databases in which the actual ISSN is stored in the SN field, you should remove SN from "link.fieldlist.p=". If you also subscribe to a database in which the ISSN or ISBN are contained in the SN field, you may wish to enable only one or the other of the databases for WebLinking, using the "link.databases.p=" entry in webspirs.user.cfg.ABI/INFORM editions (UMAF, UMAI): ISSN is in the IS field. However, it also has an SN field. For any records which do not contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.BIOETHICSLINE (BIET): ISBN is in BN field. BN should be added to "link.fieldlist.p=" ISSN is in SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.CINAHL (NURS): ISSN is in SN field and ISBN is in BN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. BN should be added to "link.fieldlist.p=".CONCISE ENGINEERING AND TECHNOLOGY INDEX (CETI): ISSN is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data.DISSERTATION ABSTRACTS (UMDA): ISSN is in the IS field. However, it also has an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.FOOD and HUMAN NUTRITION (FNUT): contains an IS field. Where there is no ISSN or ISBN, this field may contain an IPC number, which would probably lead to spurious results in your OPAC.FOOD SCIENCE and TECHNOLOGY ABSTRACTS (FSTA): ISSN or ISBN numbers are in theNU field; if you subscribe to FSTA, add NU to the list of fields in "link.fieldlist.p=".GALE ENCYCLOPEDIA OF ASSOCIATIONS (EASS): contains an SN field which contains"section number" data. EASS does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.CD ROM'S IN PRINT (GCDR): ISSN is in IS field and ISBN is in IB field. However, GCDR also has an SN field, so for any records which don't contain an ISSN or an ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.GALE DIRECTORY OF DATABASES (GDDR): includes an SN field which contains "sectionnumber" data. GDDR does not contain ISSNs or ISBNs. See above (A Matter of Fact) for solution.GPOC: has both an IS and an SN field. For any records which don't contain an ISSN or ISBN number, WebSPIRS will construct a URL using the data in the SN field. See above (A Matter of Fact) for solution.HUMAN NUTRITION (HNUT): ISSN number is in the SN field; potential problem if you subscribe to other databases with an SN field that does not contain ISSN or ISBN data. ISBN number is in the BN field. BN should be added to "link.fieldlist.p="PEST–BANK (PEST): has IS field but it does not contain ISSN/ISBN data. If you subscribe to PEST, you may want to enter the database family IDs of all of your other databases into "link.databases.p=" in order to exclude PEST from WebLinking.

What is the syntax I need to set up hotlinks within my Library Holdings message files with WinSPIRS?
The syntax needed to create a hotlink that works through WinSPIRS can be described as follows:

mmmpppmmm

Where:
mmm is information which must be included both in the front and behind the hotlink
uuu represents the URL
ppp represents the prompt text

A specific example would be:



Library owns 1976 – present A HREF="http://www.ovid.com">Click here .



Note: I placed a period "." at the end of the line to enter text after the link. Text must be present before and after the linking information.

"Bad command or file name message in DOS Window"
HCU 4.0 on Windows
This message indicates a problem with either one of the file names or the HCU installation.

1. Verify path and file names listed in your HCU MARC or ASCII conversion window are correct.

Try converting file again.

If error message still appears go to solution 2.


2. PERL 5.003 should install as part of HCU as a subdirectory in the default installation path
shown below:

C:\Program Files\SilverPlatter Information\HCU\PERL5003

It is possible HCU did not fully install.

Uninstall HCU.

Check that the HCU directory has been removed from C:\Program Files\SilverPlatter Information\ or the location to which you originally installed HCU.

If the HCU directory is still present back up any MARC, ASCII, or SPLHA.* holdings files if
they reside in the HCU directory.

Delete the HCU directory and its contents.

Run the Windows Registry Editor (regedit.exe).

In Windows select Start...Run and type regedit.

In regedit move down the following path:

HKEY_LOCAL_MACHINE
SOFTWARE
SilverPlatter Information
Holdings Conversion Utility

Remove the Holdings Conversion Utility folder.

If this folder is not present look for PERL5003 under SilverPlatter Information and remove it.

Exit the Registry Editor

Check for an HCU directory under C:\temp and remove it if present.

Reinstall HCU.

After installing check for PERL5003 under C:\Program Files\SilverPlatter Information\HCU

Try converting file again.

ERL dbserver does not display holdings messages if HCU output files are very small
Input files for conversion to "splha" format through HCU, should contain at least 50–100 records. If the HCU output files, splha.lhi, splaha.lhm, and splha.ltd, are too small there will be no holding messages displayed after placing the splha files in /sproot/cfiles.

We are a local ERL installation which uses the WinSPIRS client. Is there a way in which we can set up a link from SilverPlatter records to our web–compatible OPAC?
If you have installed the newest version of the ERL server, v.4.X, you can embed a web link within your HCU holdings message. You must adhere carefully to the instructions for doing this as described in the HCU documentation (Section 1.4.2, HCU Manual), http://www.ovid.com/site/support/system_admin_manuals.jsp
You must also be using the newest version of WinSPIRS, v.4.X.