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

Kinetica Cataloguing Standards

Kinetica Cataloguing Workflows

Cataloguing Service

Resource Sharing Interfaces Workshop, Melbourne, 5 February 2002

Background

Summary of Issues

Workshop Minutes

1. Introduction and welcome
2. Issues Panel
3. Identification of key issues
4. Reporting and discussion of issues raised by syndicate groups

APPENDIX


Background

A workshop was held on Tuesday 5 February at the Melbourne Convention Centre. The workshop follows upon a similar workshop held in 1998. It was convened by the National Library on the recommendation of the Kinetica Advisory Committee and was attended by 41 delegates representing libraries, library system vendors, book and serial suppliers and the National Library. The workshop was organised to provide an opportunity to discuss how library services are provided and how workflows have changed and are expected to change over the coming years.

Summary of Issues

  • Participants affirmed the continuing importance of maintaining the currency and coverage of the National Bibliographic Database (NBD).
  • Kinetica needs to retain its recognition and support for a diversity of workflows to meet the varying needs of Australian libraries.
  • There was general agreement for the need for closer integration of resource discovery, ordering and NBD updating where economic pressures necessitate that processes are done once.
  • A number of workflow models were discussed. There continues to be support for the 'pass through' model identified by the previous Interfaces Workshop in 1998. In this model bibliographic and order data is sent from the library supplier to the library via the NBD. The workshop suggested that an XML package might provide a suitable vehicle for this data.
  • A number of library representatives suggested a need for Z39.50 searching of book and serial vendors' databases.
  • Library vendors are developing XML for exchange of information with publishers and will be able to supply data to libraries in this format.
  • There were differing views of the value of adding 'on order' holdings to the NBD.
  • There was some support for the development of Z39.50 update for updating multiple targets; others thought the focus should be on XML packets to assist communication between vendors, Kinetica and libraries. XML was described as a more widely accepted standard than Z39.50. If this approach was adopted, Kinetica will need to support XML input/output and flow through.

Workshop Minutes

1. Introduction and welcome

Roxanne Missingham welcomed participants to the second Resource Sharing Interfaces Workshop convened by the National Library on the recommendation of the Kinetica Advisory Committee (see Appendix 1 for a list of participants). Ms Missingham noted that many issues discussed at the previous workshop in 1998, such as the need for an uploading service, have already been implemented or are no longer relevant, while other issues have emerged such as access to electronic resources. This workshop had been convened to enable an interchange of ideas between the different sectors: librarians, library system vendors and book and journal suppliers. It was an opportunity to discuss how services are provided, and how workflows have changed and are expected to change over the coming years. All sectors are affected by economic imperatives and it is essential to develop models that will assist in the streamlining of technical services. The workshop focuses on the role of Kinetica and the need to maintain the National Bibliographic Database (NBD). The workshop is also an opportunity to develop a better understanding of all the issues, which could then be built into our planning. Ms Missingham concluded that a summary of the workshop would be used as a consultative document with industry.

2. Issues panel

The workshop facilitator, Helen Livingston, introduced the speakers for each of the three sectors:

Patricia Scott, University of Adelaide representing libraries Chris Taylor of Epixtech, representing library system vendors Linda Carmichael of James Bennett's, representing library suppliers.

2.1 Patricia Scott discussed the implications of the University of Adelaide's library workflow and the technological assistance needed.

All workflows operate in an online environment where systems influence, if not dictate what is possible. Workflows are also affected by the scarcity of resources and this necessitates that processes are done once and double handling is avoided. In this context, two casualties of cost cutting have been timeliness in reporting holdings to the NBD and a lowering of standards.

Ms Scott then described the University of Adelaide's workflow. Orders are first checked in the local catalogue to avoid duplication. The priority is then to acquire the best record in the local system as soon as possible: for 85% of items, records are available in Kinetica and these are downloaded in real time at the order stage. OCLC is searched second and if a record is located in either catalogue, this saves keying and improves accuracy. Holdings are not added to Kinetica at this stage (although the library may review this in the future). Their new system allows concurrent Z39.50 searching which has advantages but as Z39.50 update to Kinetica is not available, this could not be used to add holdings. Orders are then forwarded to vendors in batches in a variety of ways (print, EDIFACT, email).

Ms Scott discussed the University of Adelaide's investigations into using vendor databases to identify and order items directly and discussed a number of issues, including:

  • vendor databases lacked good quality records available in real time;
  • searching vendor databases added complexity to the workflow (though this could be done by concurrent searches if all were available for Z39.50 searches and could be searched concurrently with Kinetica);
  • the creation of an order in their local system automatically creates a local order number. However if an item was ordered on the vendor database, the local order number would only be created once they received the confirming order record from the vendor;
  • creation of the order record locally immediately encumbers the funds allocated; and
  • without searching independently of a vendor's database, libraries may miss out on later or variant editions or related material. Kinetica gives a much wider range of information and it is therefore better to search the NBD before placing an order.

At receipt of item, workslips are printed and placed inside the item, which is placed on the shelves for review by cataloguers.

Catalogue records are reviewed with the item in hand. Headings are checked and authorities imported where needed. OCLC is used for LC authorities and Kinetica is searched (via Z39.50) for Australian authorities. The classification is checked against DDC so that related works are shelved together. If the record is of poor quality, the cataloguer will search Kinetica and possibly OCLC for a record to clone and import. Original cataloguing occurs for about 1% of items and is therefore not common in their system. For original or cloned records that are amended in any way that would be useful to the wider library community, the bibliographic record is flagged for export to Kinetica. For a crucial or complex change, the cataloguer may use the Kinetica Client to amend the record online but this is less common as the Client is more cumbersome and less familiar to use.

The cataloguing workflow concludes with the end cataloguer who completes the call number and amends the status of the record.

Ms Scott then discussed a number of issues that would assist the University of Adelaide's workflow. At the order stage, Kinetica should be a one-stop-shop to verify orders independently of vendors and obtain records. If Z39.50 update was available in Kinetica and the local ILMS, then holdings could be added to the NBD at the order stage. If libraries use a vendor database to order books, they will not want to make a second search of Kinetica to retrieve bibliographic records. If the vendor were to supply records, these need to be supplied as full records in real time (not brief records supplied in batches using FTP as some vendors currently supply); and the vendor or the library still need to supply holdings to Kinetica which may not occur until the item is received or end processing occurs.

Ms Scott discussed a flow through model where vendors supply library holdings to the NBD with the order data, which Kinetica would supply to the participating library. This raises issues for the NBD of the retention of such order data as well as commercial confidentiality and is also more likely to be done in batches rather than real time, which would cause further delays. This model also raises questions of standards. However, if vendors sent records from Kinetica this would supply records from a single source of data and therefore a single standard, which would be an advantage of the model.

At the cataloguing stage, local systems need to be more efficient in identifying unauthorised headings. The availability of Z39.50 update versus batchloading is also an issue: real time, simultaneous updating of the NBD and the local system is essential for bibliographic, holdings and authority records, with the emphasis on doing tasks once. By contrast, batchloading involves a delay during which time other libraries have made the same amendments thereby duplicating effort needlessly.

Other recommendations at the cataloguing stage included:

  • receipt of update files (bibliographic change files) from Kinetica from the order stage to the final cataloguing stage only;
  • templates that could be applied to imported or cloned records: the use of templates is widespread among ILMS suppliers but in many systems, these templates can not be applied to imported or cloned records (but only to original cataloguing of which the University of Adelaide has less than 1%) so this feature is badly under-utilised;
  • for electronic resources, the ability to save HTML records directly from MARC records;
  • automatic verification through Web Dewey of Dewey numbers added to cataloguing records;
  • use of the MFHD (ANSI Z39.71- 1999) standard for the storage, export and exchange of holdings data, particularly in the case of serials (where it allows the storage of holdings data as well as publication pattern data thereby allowing holdings to be accurately interpreted). This would allow more accurate holdings summaries that potentially could be interpreted by resource sharing modules to assist unmediated document requests;
  • global changes of bibliographic and authority records;
  • for authority control there is a need for an Australian updating or authority change service which could provide libraries with an integrated file of US and Australian headings, ensuring that those for Australian authors and specific subjects are maintained.

2.2 Chris Taylor (Epixtech) spoke from the perspective of what library system vendors would like to be able to do.

At the order discovery stage it is essential for a library to check its own database first, to ensure an item is needed. The second step would be a search of Kinetica as part of which, Kinetica would act as a broker to identify suppliers who have the item available. Having identified their preferred supplier, two alternative workflows were suggested:

  • the library would receive the bibliographic record from Kinetica with order information in the 900s, to add to their local database and have their holding added to the NBD
  • or alternatively, the library supplier would forward an XML package back to the local library system and, having accepted the order, contact Kinetica to confirm the order and obtain the record to send with the item.

So holdings could be added to Kinetica either at the order stage or the record filed in the local system and forwarded to Kinetica via Z39.50 update. This would require the ability to forward a record to a number of different sites simultaneously i.e. transmit to a local database and to other bibliographic utilities. In this context, the National Library may wish to develop software which would enable it to accept updated records information individually.

So a broadcast search to all vendors via Kinetica would enable a library to select their preferred supplier. Z39.50 extended services would enable MARC records to be forwarded by FTP, with the emphasis on doing tasks once and avoiding double handling.

In response to a question about selecting a library supplier following the broadcast search to Kinetica, Chris Taylor replied that the functionality already exists to canvas library suppliers through a single search. However library suppliers would have to buy in to the arrangement and there may be issues such as commercial-in-confidence to be worked through. The issue is political not technical and would require a change of approach. In conclusion, Helen Livingston referred back to the supply of an XML package to libraries and described it as a very different way of operating.

2.3 Linda Carmichael (James Bennett) began by addressing some of the issues raised in Chris Taylor's paper. Ms Carmichael noted that library suppliers are already using XML for exchanging data with publishers, enabling them to create a single record with all the information needed. However XML records have not been transferred to libraries as yet.

On the suggestion of interrogating vendor databases via a Z39.50 broadcast search, Ms Carmichael thought this was probably not the most appropriate prodedure. As an example, Bennett's titles database (designed as an interface with publishers) was not stored in MARC format and would require a lot of work before it could be interrogated via such searches. Library suppliers generally prefer to load order information and supply data to individual libraries in batch mode. However as library suppliers are developing XML for exchange of information with publishers, they will be able to supply data in this format.

Ms Carmichael referred to the inconsistencies in order information received from libraries using different library systems. Currently data could be supplied in around 40 different ways. Library system vendors move from one standard to another, each of which picks up more or less information, which can cause problems. Library suppliers would like more consistency from library system vendors, ideally a single standard for order information. On the other hand, library suppliers would like to transmit more information to customers but, again, there are enormous variations in the technology available to customers and in the information they want to use.

For James Bennett, original cataloguing is done directly onto Kinetica as Bennetts does not maintain a database of MARC records. Library suppliers would like Kinetica to provide a different way of identifying what work they do, to provide more management reporting of their original cataloguing, updated records and addition of holdings. At present, library suppliers sign on to Kinetica using the customer's signon and the customer is identified in the 040 tag. Library suppliers want some attribution of who actually did the work, such as adding a library's holding. On the issue of prepublication information, library suppliers would like to send this data to Kinetica so it would be in the public domain much earlier than at present.

In response to a question about the poor quality of the records libraries often receive from library suppliers, Ms Carmichael replied that, while library suppliers have bibliographic information in their databases, this is not used as a catalogue and the information is gathered from many sources outside of bibliographic utilities. Vendors can usually provide a brief, automatically generated MARC 21 record at point of order but will not be able to supply all the information libraries require primarily because title information is gathered without the book in hand. On a positive note, Bennetts is currently redeveloping its titles database and looking at extracting data in XML and then mapping it to MARC21. This will potentially allow for publishing related information, currently not stored with the title record, to be transmitted automatically in the MARC21 record.

In further discussion the point was made that library suppliers operate between libraries on the one hand and publishers on the other and have to create data that can be communicated to both. Publishing information can be inconsistent up to the time that the book is actually published, with changes in ISBNs or even re-using previously issued ISBNs. To manage this, the book supplier usually has a single record for each ISBN. This differs from libraries where a library catalogue may have a single bibliographic record for a title with multiple ISBNs. For library management systems exporting title information for order purposes, it is usually the first ISBN 020 tag that is exported, even if it is not the ISBN for the edition that the library really wants to order (eg paperback ISBN when the hardback edition is really required). If would assist library suppliers if the library could nominate the actual 020 tag that is to be exported. Libraries want quality records from vendors at point of order but book vendors are not in the business of creating such records.

3. Identification of key issues

Following on from the issues panel, Helen Livingston summarised the key issues as follows:

3.1 Resource discovery issues: there is a real need to streamline the resource discovery processes through local library systems and vendor systems.

3.2 Having identified the items libraries want to order, many other steps need to occur:

  • an order needs to be placed in a way that updates the local library system immediately and links the unique order number with the bibliographic information so avoiding duplicate orders;
  • records need to match so the local library system can recognise if a duplicate is being ordered;
  • the bulk of this work needs to be undertaken by the most junior staff; and
  • libraries need to update the NBD on their holdings. This raises the question of what information is appropriate to record on the NBD and whether libraries should update the NBD with items on order.

3.3 Once an item has been received by a library

  • the NBD needs to be updated;
  • libraries may or may not want to receive update files (bibliographic change files) from Kinetica;
  • there are issues associated with reporting holdings to more than one place;
  • a separate process is required for cancellation of orders;
  • interfaces have to operate easily and efficiently; and
  • all parties need to reassess why they want to know specific data and what they really need to know.

Ms Livingston summarised the arrangements for the workshop session to follow the afternoon tea break and encouraged all groups to discuss workflows and to identify what needs to occur in all sectors to improve those workflows.

4. Reporting and discussion of issues raised by syndicate groups

Following on from the workshop, each group reported back on the issues raised in discussion.

4.1 Group 1

  • The priority is to forward information to Kinetica and this should be quality records of a realistic standard.
  • Promotion of the addition of collections to the NBD to all libraries - small libraries are essential members of the Australian library network and their collections should be on the NBD.
  • Access to Kinetica needs consortia/ site licence/ better charging models especially to encourage small libraries to join.
  • Interlibrary loans/ resource sharing is another area to be explored.
  • Prepublication information - book vendors can supply Australian prepublication information but there may be commercial issues.
  • Publishers use title information for very different purposes to libraries.
  • The linking from records to vendors needs discussion by vendors. It may be feasible to download records from Kinetica and send them to vendors for ordering.
  • Lots of workflow modelling is needed.
  • Emphasised the importance of updating holdings quickly.
  • End user access to Kinetica is an important area for development.
  • And it is still a magical, wonderful service.

4.2 Group 2

  • Workflows - need to model 'ideal' workflows for different types of materials (e.g. print vs. electronic, Australian vs. overseas) and/or for different types of libraries.
  • There is a need to identify what is the fundamental record that can be used by publishers, vendors, suppliers and libraries.
  • A Z39.50 Client that can upload records to Kinetica should be developed but who should develop and distribute it? This would allow holdings to be updated.
  • Is Z39.50 being replaced by XML? If so, support is needed for XML input/ output and flow through in Kinetica.
  • Should Kinetica be the first stop or the last i.e. where does Kinetica fit into the workflow? There was support for Kinetica as a one-stop-shop for item identification and availability and also for the idea of having to input order data once and that data flow to supplier, NBD and the library's local system. Alternatively, order data could be transmitted from the vendor to the library via Kinetica. Kinetica has a role in improving connectivity.
  • There is a need to identify the business needs for each part of the industry.
  • What would we want in an ideal world? No duplication of effort; all systems standardised; a painless and less expensive process for adding data to Kinetica and for vendors to do everything that libraries need.

4.3 Group 3

This group began with a discussion of Chris Taylor's proposed workflow i.e. local duplicate checking followed by a single, broadcast search of Kinetica and vendor databases. They concluded that this model would need holdings reported to the NBD while items were on order, which brought up the issue of standards:

  • Z39.50 update would allow the updating of multiple targets. Kinetica could develop this or the software could be developed by a third party vendor.
  • EDI - is it specific enough for the library industry or is something more specific needed? Alternatively should we be looking at XML packets to assist communication between vendors, Kinetica and libraries?
  • Further discussion of workflows: discussion with vendors probably requires two transactions, not one; vendor or local library system can forwarded 'on order' holdings to Kinetica; MARC records could be supplied from the vendor or from Kinetica but how to manage the remainder of original cataloguing?
  • Workflow issues between book and serial suppliers and Kinetica: how do vendors know what is on Kinetica and obtain it for their local database? They could forward a file of ISBNs to Kinetica and Kinetica send back the best available records. In this model there are issues in defining 'best' and this would ultimately have to be done by a person. Secondly vendors need better arrangements for supplying information to the NBD, including prepublication data and to be able to add bibliographic data under their own signon (which would be a form of advertisement and so may encourage them to add more). Vendors also need to be able to add holdings for groups of clients rather than individually.
  • Australian authority records - there are issues in defining Australian authority records and this is being addressed as part of the Expert Advisory Group on Authorities.
  • Support for the MARC holdings format (Z39.71) and the more accurate holdings summaries that it could provide would be especially important in assisting unmediated interlibrary loans.
  • Briefly discussed confidentiality issues between supplier and client - no conclusions as to how important this may be, or how such issues could be resolved if order data is supplied with MARC records to NBD.

4.4 Group 4

Group 4 considered a range of workflow issues, including the following.

  • Different workflows are needed for different types of materials eg monographs and serials. Z39.50 searching does not work well for searching serials, which have to be searched through the Kinetica Client. Perhaps serials could be put into a separate database on the NBD?
  • Different workflows are also needed for large and small libraries and Kinetica needs to retain its recognition and support for a diversity of workflows.
  • For those contributing data online, the Kinetica Client is too slow. If it were more efficient then libraries would be more willing to clean up problem records.
  • If holdings were added to the NBD at the 'on order' stage then cancellations could be removed by batch deletion through BatchLink.
  • Ø XML vs. Z39.50 update: Z39.50 update would overcome problems with the Client and delays from using batchloading. Alternatively XML packages could be extended throughout the process, allowing staff to search Kinetica from their library management systems. XML is a more widely accepted standard, used by publishers and accepted outside the library world and was described as better for searching. However for both XML and Z39.50 searching and update, smaller libraries do not have the funds to access this technology.

4.5 Group 5

Group 5 looked at organizational boundaries, that there are different needs and priorities for all groups in the chain. This group considered the following.

  • What are we trying to achieve? There is no consensus and different views need to be brought together.
  • Many of the issues are about library workflows (i.e. not about Kinetica).
  • The best source of data may not be part of the same chain.
  • What is the centre of the data flow? If we focus on forwarding bibliographic data to libraries, then the vendor is at the centre. Could searching and placing an order on a vendor's database trigger a Kinetica search with Kinetica supplying the record to the library? Where data is not on Kinetica, how should it be obtained? On the other hand, vendors may be reluctant to open up automated/ distributed searching due to pressures on availability and price.
  • Automating searching and matching leads to a trade off in quality.
  • Where does Kinetica fit into workflows? At both the beginning and the end as data flows in multiple directions.

4.6 Group 6

This Group focused on standards issues across the workflow, specifically:

  • Kinetica is a standard database available to all libraries. Yet there are quality control issues for bibliographic records including inaccurate information ("too much rubbish in bad records") and minor differences leading to the creation of duplicate records. It was also thought not useful to add 'on order' data to Kinetica as the data could change.
  • There is a lack of standards in workflows. The retention of many local practices is generating much work in order to maintain them.
  • Workflow issues between book and serial suppliers and Kinetica: at present vendors need to access Kinetica each time they wish to download a record for a library. It was noted that vendors make no money from supplying records to libraries but only do so to sell books. Vendors need more efficient workflows and equality of charging with libraries. Kinetica was also described as too slow.
  • The focus should be on providing information to the end user.

Helen Livingston concluded the workshop by thanking all participants and stating that the information gathered would be used to prepare a paper identifying the actions to be addressed.

APPENDIX 1

LIST OF INVITED ATTENDEES TO PARTICIPATE IN THE RESOURCE SHARING INTERFACES WORKSHOP

Anderson, Craig
Chief Executive Officer
YARRA PLENTY REGIONAL LIBRARY

Beaumont, Anne
Applications Support Coordinator
Technology Services
STATE LIBRARY OF VICTORIA

Bracher, Elizabeth
Director
BIBLIOTEK

Brodie, Maxine
Director
Strategy and CIO STATE LIBRARY OF NSW

Carmichael, Linda
Technical Services Coordinator
JAMES BENNETT Pty Ltd

Chan, Iming (unable to attend)
Manager
TRANSLIB Pty Ltd

Cherrett, Carolyn
Marketing Manager
EBSCO Australia

Cohn, Carolyn
Director
BLACKWELL

Cook, Mary
BIBLIOTEK

Crawford, Jock
General Manager
FRETWELL-DOWNING INFORMATICS

Diradji, Din
Database Maintenance Supervisor
MONASH UNIVERSITY, Clayton Campus

Franks, Murray
Manager
XJs Pty Ltd

Gatenby, Pam (unable to attend)
Assistant Director-General Collections Management
NATIONAL LIBRARY OF AUSTRALIA

Gawne, Jennifer
CAVAL

Goldner, Matt
FRETWELL-DOWNING INFORMATICS

Henczel, Sue
Education & Business Development Manager
CAVAL

Hill, Claire
Product Manager
ENDEAVOUR INFORMATION SYSTEMS INC

Horn, Mac
INNOVATIVE INTERFACES INC.

Livingston, Helen
Acting University Librarian
DEAKIN UNIVERSITY

Maslen, Jan
Associate Librarian (Technical Services)
TROBE UNIVERSITY

Neil, Alison
Product Manager
GEAC COMPUTERS Pty Ltd

Novakov, Nada
Sales and Marketing Manager
JAMES BENNETT Pty Ltd

Roberts, Susie
Manager Library Services Treasurer
MPLA CAMPBELLTOWN CITY COUNCIL

Rousseau, Glenda
EX LIBRIS (Aust) Pty Ltd

Sadler, Roderick
Senior Cataloguer
NATIONAL INFORMATION AND LIBRARY SERVICE

Scott, Pat
Bibliographic Services Librarian
UNIVERSITY OF ADELAIDE

Smith, Jack
Library Systems Development Manager
SANDERSON

Tassell, Elaine
CSIRO Information Management

Taylor, Chris
EPIXTECH Pty Ltd

Wade, Rona
UNILINC

Wakely, Trevor (unable to attend)
Technical Services Librarian
NATIONAL METEOROLOGICAL LIBRARY

Walkden, Helen
Manager Technical Services
PLAIN CENTRAL SERVICES

Walters, Stephen
DLS Australia

Woodley, Leonard
Managing Director
ALS LIBRARY SERVICES Pty Ltd

Wright, Robyn
Account Manager
SIRSI Australia