National Library of Australia
Kinetica Kinetica
This page is an archive page and will no longer be updated.

Annual Users' Meeting 2000

Cooperative cataloguing in Australia:
issues that help or hinder it.

 

Most of us would be aware that the stated focus of the new Kinetica paradigm is on holdings and resource sharing, but holdings for the most part have come as an integral part of the process of cataloguing, cooperatively. Although it is relatively easy for libraries to structure their contribution to the NBD around contributing holdings alone, it is unwise to think that resource sharing is more important in this cooperative venture than the need to contribute and use the bibliographic data that has been traditionally part of the NBD.

A significant achievement of ABN was the quality of the catalogue records. The ability of libraries to improve records to a defined standard has sheltered us from the full impact of declining standards in international cataloguing as major agencies cut costs. Although we sometimes grumbled about records from particular sources that fell short of the standard we felt was necessary for our users, for the most part substandard copy for commonly purchased items was corrected promptly, usually within about two weeks of it hitting the NBD. We have been regularly using copy from two other sources for more than five year. The number of basic changes we are having to make to records from Kinetica seems to be catching up with the number of changes required on records from OCLC or RLIN. The pity of it is that most of us are probably all individually fixing up the same records.

This lack of timely, quality maintenance on the NBD will undoubtedly affect interlibrary loans too. If typographical errors in titles or ISBNs, or non preferred form of names for authors are commonly coming through onto the database, and not being corrected expeditiously, the chances of missing items and holdings are greater.

For our part this is mostly our own fault. It is one of the downsides of batch loading. This is the path that we chose to take. However all of us are required to look hard at our own productivity, and choose the most streamlined approach to our workflow, so to some extent we cannot help the choices we have made.

Kinetica now offers two main ways to update the bibliographic database, batch loading being an option which was really only reserved for retrospective files with ABN. If we choose this path we need to ensure that it has the best outcome for all. This really means doing what we can to ensure that both original cataloguing and significant improvements to catalogue copy gets to the NBD as quickly and as accurately as possible.

If we continue to batch load, we need to make this as slick as possible at both the Kinetica and the local end. It would be good to air the topic of better development and support for more regular batch loading. However batch loading as it stands at present is also very much a blunt instrument in getting data to the NBD.

Some suggestions to improve batch loading:

1. Better and detailed reports after batch loads would be welcomed. We do not know which records have not been accepted onto the NBD when a few are rejected in each batch, so we cannot then go through the client and alter/add them directly, or correct the records to resubmit. We also cannot tell which records that we intended to update existing records (as they include a Kinetica number) have actually generated new records on the NBD instead. If we are batch loading it is essential at least to know which records have been rejected, so that we can either modify our rejects to resubmit, or alternatively update them online, if we are able. Otherwise we cannot completely fulfil our cooperative role.

2. If a file was returned that were able to report our local record number and Kinetica number in sections, new records added, Kinetica records amended and a report of rejects with our local record number and perhaps the system message of reason for rejection we could more easily follow up on any mishaps. If we have the system capability we would also be able to add the Kinetica RID into our local records, so that future changes would be easier to match with the correct NBD record.

3. An obvious place to put batch files on a Kinetica server, so that we can see what we have sent at what date, and when they were processed could be useful to help us keep track of what we last sent, and possibly make it easier for the Kinetica team to process. At present they have to go to various different sites and collect data when notified. If we could also see what was in the pipeline generally it would enable us to decide whether to hold off some original cataloguing for a particular group of Australian publications if a file from that agency were in the queue.

4. An understanding of what factors are used by Kinetica to make a match, or create a new record would be useful. If there are mechanisms that would enable us to be confident that a revised record would indeed match and overlay an existing record, or what parts of it would "merge" it would be useful to know in detail. If we knew what types of data would be better to add on-line to avoid duplicate/conflicting messages we could oblige when the need arose.

5. We are aware of many duplicates on the NBD. Although a large percentage that we notice are ex-ABN, there is a growing feeling that many new duplicates are being created. Perhaps the fact that records are more easily found with the new software is making the issue look worse than it is. We were concerned that imperfect batch load matching may be contributing to the problem. However many seem to relate to international loads. Examples to hand suggest differences in pagination between a CIP and a full record, or choice of places of publication recorded in the "260 a" according to origin of cataloguing, or even variations in the fullness of the authors name from different agencies are causing many duplicates. Perhaps this has always been a problem, and is difficult to resolve. But perhaps we are not taking such care to fix up what we find or report duplicates as promptly as we did before.

6. We should try to work towards an agreement between the libraries that have the choice of working online or off-line, as to what corrections are highly desirable to be done online, immediately, and what can wait for later batch update. Not all would want or be able to support this, but such an agreement would give us a feel when restructuring workflows and policies, and give us something to point to. One suggestion could be that fixing typographical errors in key access points, such as title or ISBN should be done online (and are relatively quick and inexpensive to do), whereas additional headings could wait for a batch update. Reporting recently added duplicates promptly, perhaps with an online form to help, or a special email address would be another useful feature to help keep the database clean.

7. Investigation should be undertaken to examine those factors that have encouraged libraries to delay their cataloguing contribution or give up altogether.

  • Our cataloguers resented the software that they were required to use in the Amicus Client, especially as we had been successfully using a very quick, flexible, interactive Windows based product locally for some time in retrospective upgrading. Improvements in the AMICUS software would assist this issue, especially concerning slowness in transferring items between headings, and author/title headings, and having to go in and out of indexes to check headings rather than the ABN "check" type of facility.
  • However batch loading offered us the ability to focus on our own catalogue, which had been neglected in favour of ABN, almost the sole environment in which our cataloguers had worked. Improvements to the local catalogue since the change to Kinetica have been significant, and we wouldn’t want to lose the advantage gained, probably even if the AMICUS client became the most flexible tool imaginable. (Although I am sure we would weigh the pros and cons)
  • It is clear from earlier meetings that some other libraries have been unable to use the Kinetica client because of firewall difficulties. It would be interesting to survey what outcomes have eventuated from these difficulties.

Alternatives to batch loading

Alternatives to batch loading that require little systems staff intervention and also allow us to interact well in real time with our local system and the NBD are options that need to be considered seriously. Z39.50 still offers this promise, but may not be the only option.

Some indication of the priority of Z39.50 UCP as a development for AMICUS would be helpful. However all Z39.50 solutions rely on the lowest common denominator in terms of server and client.

It would be helpful if Kinetica were able to suggest suitable Z39.50 clients that could be purchased by users to integrate with their own ILMS, or a list of ILMS vendors that have developed or are in the process of developing Z39.50 UCP. Or alternatively sponsor a local vendor to produce a product that we could "add on". The National Library organised a very useful meeting with representatives of libraries and vendors before Kinetica began to look at workflow models. Maybe it is time to explore possibilities with software providers again.

Keeping headings in our local database and Kinetica in sync

An ability to download Kinetica authorities or capture them through Z39.50 would encourage work on Australian authorities even for those libraries that batch load bibliographic files. Keeping headings clean on the NBD would make it easier for us to avoid duplication of headings on our local catalogue, and help to keep headings on records we export to the NBD in batch mode in sync.

Similarly making LC authorities available for downloading or Z39.50 capture would ensure that our non-Australian headings are up-to-date.

Allowing us to easily augment Kinetica from other sources

It would be good to know if there could be developments to allow us to use the Z39.50 client capabilities (if they exist) of the AMICUS Client to point to another database eg. RLIN, or a local catalogue and directly capture a record to the NBD, and not to have to save it line by line would also be a way to make it easier for libraries to work constructively with the NBD. This would be good for both bibliographic and authority records. If we were able to point it at our own database it would provide yet another possible alternative to batch loading for some.

Standards

With ABN we had a cataloguing manual and a Standards Committee. We knew what was acceptable and what we should do to provide quality records. (Even if we didn't always follow them perfectly). However cataloguing rules are in flux. It is very difficult to know what we should be accepting or amending in a number of areas. And we are now sourcing records from a wider variety of databases, with their own rules.

I am aware that work is being done on Electronic resource cataloguing, and hopefully that Expert group will provide us with either a new chapter of the manual, or commentary on an existing manual (eg. N. Olson, Conser etc). In the current state of change we need this more than ever.

The mounting of the original ABN Cataloguing manual on the web is a great start, but it is fossilised, when it needs to be gradually updated by documents relating to cataloguing in the Kinetica environment, linked to the pages of this original manual, as topics are brought into line with current trends eg. policy regarding F.2.1 Romanisation of Chinese needs to be updated in light of recent LC changes. (I note that a CJK user has added this particular issue as an item to the ANCJK User meeting agenda.)

It has been a major focus of the Kinetica team to get a great deal of documentation such as the searching guides and edit manuals on the web, so making them widely and easily available to us all. I applaud this responsiveness, and the willingness to allow other libraries to also post some of their helpful documentation. We just need to go beyond the technical "how to" documentation, and get back on track with providing guidance among ourselves. This helps us leverage our skills in interpreting international trends and ensure useful cataloguing decisions are easily made.

Patricia Scott
18th August, 2000

*Some statistical indicators.
While it is difficult to say that we are comparing like with like, as we only search other databases for material not found on the NBD, therefore potentially more esoteric, difficult cataloguing the total percentage of changes to copy cataloguing has grown from 29.6% in 1997 to 36% in 1999. At the same time, original cataloguing has grown from 13% in 1997 to 18% in 1999, despite the fact that we are heavily geared to do copy cataloguing first. The amount we are changing online on the NBD has remained reasonably constant : 8% in 1997, 10.7% in 1998, 9.2% in 1999. We are changing more records locally, and have increased our use of records from OCLC and RLIN from 4.2% of the total in 1997 to 8% in 1999.

Draft agenda Kinetica Annual Users' Meeting 2000