"The time is right to re-invent the online union catalog in the image it needs to carry it into the 21^st century … we think the resulting system is going to be more than just another service from OCLC. We think it will be a major factor in how libraries use the Web and electronic resources." (Hickey, Childress, & Watson, 1999: 30-31). "CORC has the potential to be one of the most significant developments in modern librarianship this century – perhaps second only to the development of OCLC itself." ( Irvine, 1999: 40)

The hype surrounding the introduction of CORC (2000) has certainly been compelling, and like most libraries the UNSW Library is looking at all options for improving access to the increasing number of Web resources. To check the claims concerning CORC, the Library decided at the end of 1999 to join the project to see for itself what the new system had to offer. We realised that the system was evolving and was a work-in-progress, but also saw an opportunity to have input into its development.

Why did we bother joining the OCLC system when we have our own national utility, Kinetica, and many other Web search tools? Like most libraries we do not believe that our Web access tools are optimum. Searchers still have to use several routes to find material – the catalogue, Web lists, subject gateways etc. We are, therefore, open to experimenting with various approaches. Librarians everywhere are struggling with the question of how to best point to Web-based resources. No one can be sure whether one approach will predominate in the future, or whether we are going to continue with the current range of access alternatives. Ideally librarians will contribute to the development of a single interface for information seekers. Participating in the project, we thought, would allow us to investigate OCLC’s claims for the new libra rians’ search engine, and to assess the cataloguing system and pathfinder creation tools. We needed to consider more closely whether the catalogue will become the one-stop shop or whether another integrative mechanism will be needed.

Attractive features of CORC were the Dublin Core/MARC crosswalk, the automatic suggestions for keywords and Dewey numbers, the built-in authority system, the export options in MARC or Dublin Core, the promised maintenance reports and the pathfinder creation tools. Using the system has provided an opportunity for the cataloguers and the Library to learn more about metadata through practice, and to experiment with exporting records for Web resources (in addition to those for electronic journals) into the catalogue.

 

The main reason, as mentioned ab ove, is to participate in a practical, hands-on way in a system, which uses new technology to create its data and to assist Web access. It is not possible for libraries to go it alone in developing systems and to keep up with the latest research in our arena. A global cooperative project enables the participants to share the costs and to collaborate and network as the new field is opening up.

The UNSW Library aims to fully catalogue electronic journals and databases. We do this on Kinetica by creating new bibliographic records or adding details and holdings to records for print versions. The same journals and databases also appear in lists on our Web site. Our Web site also provides links to search engines that are often used for finding free non-core resources. Many of these free resources and sites are valuable to our users. By exporting bibliographic records from CORC into the catalogue we can see how useful this approach may be. An alternative option for CORC libr aries is to leave records in CORC and point users to that database.

Another motivation was to increase our expertise in metadata. The library has already been using metadata – mainly Dublin Core, and saw this project as a way of extending the experience of cataloguers and other librarians. Our Web site includes metadata embedded in Web pages, and we do this by using our own template (1999) adapted from the Nordic project’s online template. We also specified the metadata for automatic generation in the Australian Digital Theses Program (2000), and have added metadata to the MetaChem (2000) and AVEL (2000) subject gateways. CORC involvement seemed a natural progression following these first experiences with Dublin Core. Metadata is becoming increasingly important to librarians and we are aware that it is crucial to be up-to-date and to involve as many of our staff as possible.

We are also just about to embark on the search for a new in-house system. We thought t hat an examination of the features of CORC would be useful in this quest. By actually using leading edge technology we would have the opportunity to see at first hand which features should be essential or desirable for a new in-house system. Of particular importance is the fact that the Library has no in-house cataloguing module, and does all its cataloguing online into Kinetica. We were therefore especially keen to test the value of these new cataloguing tools offered by OCLC.

In summary, the search for best ways to access Web-based resources, the need to examine features for our new in-house system and the opportunity to work for free on CORC using the new technology, with the opportunity to learn more about Dublin Core, HTML, XML and automated features were the main reasons we joined CORC.

During our involvement we needed to assess what the system had to offer our staff and users – i.e. to cut through the promises to determine the real benefits. From July 2000 the service would be full-fee paying so we also needed to assess its cost effectiveness. At the end of June 200 we concluded that we needed some more time and experience to assess the full range of system features.

Although when we joined up the requirement was for a half of a FTE staff to work with CORC, this had been dropped by the time we were signed up, so fell away as an issue. We then used the system as time permitted.

Prior to the CORC Project our involvement with OCLC had been minimal. We use FirstSearch for interlibrary loans and have signed up as a partial user through Kinetica. As partial users of OCLC we suspected that we might have to pay more for using CORC once fees were introduced. When the fee schedule was released at the end of the project this was confirmed with searching and exporting of CORC being more expensive. Details follow in the "How Well Did CORC Shape Up?" section.

Our cataloguing, in particular, and to some extent, our ILL functions depend on Kinetica. Many of the Library’s operations and workflows are centred on the Australian system. We are involved through local and national user meetings and have easy telephone access to staff whom we know and see from time to time. There is a feeling of involvement and control in the use of a local system. OCLC on the other hand is largely a US system in spite of its international spread. There may also be less opportunity to influence developments to the same extent, and definitely less chance to attend user meetings. Communication with CORC has been via email, with an overnight response.

Another factor is the developing, evolving nature of CORC. It is unfinished and software has been continuously upgraded since 1999. The crosswalk between Dublin Core and MARC is still being developed, for example.

While we saw the project as an opportunity to gain experience with metadata, X ML, HTML, the Dublin Core/MARC conversion, automated features, exported bibliographic records etc., there is very little in the way of tangible benefits to our user community at the moment. The involvement must be seen as more of an investment, a learning experience, an experiment, at this stage. The more records we import into the catalogue, the greater the chance there will be to assess the value of their inclusion.

There were also very few other Australian participants with whom to compare notes. The group is small– the National Library, Melbourne University, Monash University, Adelaide University and Flinders University (there are probably others) and the Australian experiences have been limited.

Exporting records from CORC meant that another workflow had to be established and technical assistance procured. Systems staff are always too few relative to the demands made on them, and as project leader I had to ensure that our workflow was not too demanding of cataloguing and programming staff. During the project 40 records were created and 20 were imported into the catalogue. They were manually exported, loaded into the catalogue and linked to the WWW.

The Metadata Coordinator led the project together with a small steering committee. One serial original cataloguer, 2 monograph original cataloguers and 2 reference librarians volunteered to work on the project as their other duties permitted.

The project commenced soon after an excellent four and a half-hour teleconference training session in March 2000. Bradley Watson from OCLC talked the roomful of librarians at our end through the whole system – administrative features, and the cataloguing and pathfinder creation systems. The librarians then familiarised themselves with the system prior to starting work. The project leader met with the two groups cataloguing and pathfinder creation, separately on a weekly bas is. With only a few months remaining before the end of the project period they aimed to work through all the features of both aspects of the system as quickly as possible.

Very early on the UNSW pathfinder team realised that the pathfinder creation tools were not an improvement on their current tools. The pathfinder component of the trial therefore finished early at the end of May while the cataloguers continued their work right up to the end of June project deadline

Several libraries have reported their project work on their Web sites so the UNSW Library followed their lead and set up a page (2000) to link to reports, and MARC and HTML exported records. We have found this page worthwhile for the information of our own staff and for other libraries.

To wind up our involvement in the project the Steering Committee and the practitioners met with the Principal Librarian to demonstrate the system, to highlight the issues and to formulate action plans.

The following issues were identified for investigation:

 

CORC pathfinders and catalogue records provide two different approaches to accessing Web-based resources. The UNSW Library has been using both already. Electronic journals and databases are fully catalogued and appear in Kinetica and in the local catalogue. The reference librarians have also created many subject guides (the equivalent of pathfinders) which appear on the special library Web sites. These two methods of pointing to resources reflect the fact that librarians are attempting to provide a variety of solutions to Web access.

In the Project we had two groups of librarians working on these separate areas - the reference librarians testing the pathfinders, with the cataloguers doing the cataloguing. The toolkits are quite different and we evaluated the two separately. Resources in the pathfinders do not have to be catalogued into the resource catalogue, and we made no attempt to specifically catalogue material listed in out test pathfinders.

This component of the system offers a growing database of over 1,000 pathfinders as well as tools to create pathf inders by 3 methods harvesting, cloning and creating from a blank template. CORC is planning in the future to notify libraries of broken links, thereby providing a valuable maintenance function. Libraries will still have to do their own maintenance work to mend links.

Resources listed in pathfinders may be Web-based or traditional printed material. The Web resources listed may or may not be included in the CORC catalogue. Libraries can also indicate whether resources are available only to their community via site licenses or password access. Pathfinders can include dynamic searches to provide users with up-to-date, new material added to the CORC resource catalogue since the initial pathfinder creation.

The service sounds impressive in theory with all these features. Yet the CORC pathfinders are not easy to create, and the system at this stage is not recommended by the UNSW reference librarians who worked during the project. The process is slow and clumsy and the librarians much prefer the tools they are already using, mainly Dreamweaver.

The cataloguing shortcuts make it possible to create a record on the Web without even logging into the CORC system, provided a password is entered. Data can also be harvested from selected Web sites. These two features make the record creation process very fast. This is not the end of the story, however, since careful editing is still required. The other automatic tools which suggest key words, and Dewey numbers also necessitate careful scrutiny so that any misleading results are eliminated. Also, now following the introduction of fees, libraries must subscribe to Web Dewey to benefit from Dewey and subject heading suggestions.

The cataloguers found the authority system quite good. Once a head is checked and conforms to LC headings it appears in blue and is underlined, indicating a hyperlink to the authority system

Constant data that may be requ ired for a number of records can be inserted to save coding time although the cataloguers found that cloning records was usually easier for their purposes.

The database of over 350, 000 records, initially seeded from the Intercat and NetFirst projects is rapidly increasing, and is a very valuable resource in itself. The quality of the bibliographic records varies considerably. Many are very full MARC records while others are the result of harvesting Web sites and appear not to have been edited. This does require librarians using the records to be vigilant and take time to amend records to desired standards. CORC Dublin Core input standards are being drawn up, but generally Dublin Core records converted to MARC are tagged Level 3 which is lower than OCLC minimal level (K) records. After June 30 2000 CORC records started being included into WorldCat, and WorldCat records with 856 fields are incorporated into CORC. With standards for Dublin Core less defined and much more flexibility in that format, cataloguers working with the Dublin Core format in CORC are less constrained and are more likely to leave the default data in place if they are using the harvesting feature.

The UNSW cataloguers concentrated on editing the records in MARC during the project because of their expertise in this area, and because we were importing the records into our local catalogue. They did, however, become familiar with Dublin Core as they went along, by comparing that version with the MARC record. Although the crosswalk between Dublin Core and MARC is a considerable achievement, it is still being fine-tuned. Because of the differences in the formats it will never be a perfect conversion. It is quite difficult to create "perfect" versions of each format with a change in one often resulting in an error in the other. This is the nature of such a crosswalk, however, and not just a feature of CORC.

The CORC fee schedule was released in May 200 0. For partial users fees are higher than for members who catalogue all their material on, or batch load all their bibliographic records into WorldCat. Like most Australian libraries we are not in a position to become full members at this stage. As a result there are higher fees for searching (US$0.95 compared to US$0.75 per search) and exporting (US$0.30 compared to US$0.20 per record). Partial users also miss out on credits for records created (US$2.50) which puts them at a distinct disadvantage. The CORC system is also a better proposition for current OCLC users who have their workflows already established with OCLC. To be more precise regarding costs, an estimated A$4.00 worth of searches (3) per record plus A$0.50 per record exported brings CORC costs to an estimated A$4.50 per title.

There are also staff costs, with technical staff needing to be involved and an additional workflow required. Cataloguing staff are also investing time into the project and are, of course, still required to cope with a continued influx of traditional print material. During the second half of 2000, however, the CORC records began to be automatically exported to Kinetica and thus became part of our regular workflow. This saved considerable time.

In an era of change and with the need to experiment, the UNSW Library decided to pay these costs, at least in the short term – initially until the end of 2000.

As mentioned above, the main immediate benefit to the UNSW Library is the inclusion of the bibliographic records for selected Web sites/resources in the local catalogue. By continuing cataloguing beyond the project period the cataloguers had an opportunity to learn the system more thoroughly, to consolidate their metadata expertise and to further assess the features for our new in-house system.

Although selection of resources was not a main concern during the projec t itself due to the shortage of time and the need to assess system features across a range of material, it is obviously a major issue for libraries. Do they use the same selection criteria and methods for Web resources that are free as they do for purchased print and electronic material? What will the priorities be? Initially at the end of the project, it was thought that the resources in the existing UNSW Web-based subject guides would be catalogued since a process of selection by reference specialists had already taken place. Our special librarians, however, preferred to select material for CORC cataloguing and send lists to the cataloguers.

The CORC project team met with the heads of the special libraries and the reference librarians at the conclusion of the initial project period regarding the continuation of the CORC involvement. Discussions centred on material to be catalogued, the role of the catalogue and how to best assist access to the Web-based resources. It is hoped that the CORC work and the practical results of having records imported and appearing in the catalogue will help identify issues that we have to address in the coming months. There are a variety of views within the library regarding the role of the catalogue and how we should best guide our user community.

We have yet to examine how useful the CORC Dublin Core/MARC might be for the Australian Digital Theses Program. Currently most ADT libraries are fully cataloguing print versions of theses in MARC, and as an extra process enter some new theses into the ADT database. It may be worthwhile to use the metadata automatically created through thesis deposition to create a basic CORC record (through the harvesting process), export that record in MARC and have cataloguers upgrade the record to a suitable level for the catalogue.

An Australian version of CORC might appeal to, and be more relevant to, Australian libraries, which now focus much of their work ar ound Kinetica. It will be interesting to see how Kinetica handle metadata in the future. Will any CORC-like features become available? Australian libraries might be interested in a cost-effective local solution. Very few Australian librarian libraries are heavily involved with OCLC in spite of its strong research arm and the new systems being offered.

As mentioned above we will use information gleaned through our CORC involvement to assist our selection of a new in-house system. For example, how important will formats other than MARC be? Do we need to be able to accommodate non-MARC metadata as well as MARC?

CORC needs to be seen in terms of what it represents for libraries and similar organisations and for individuals. It is in this context that its full value can be assessed. CORC is trying new ways of serving users – mainly in the cataloguing area, though the pathfinder tools may be useful for some organisat ions. What it is doing is setting up Dublin Core as an option for cataloguing. Other formats may follow. Few people are saying that MARC is going to disappear tomorrow. It has served us well as a tool to describe resources. CORC allows libraries to work in Dublin Core if they choose to do so. Perhaps Dublin Core will be sufficient for some Web-based resources, and CORC Dublin Core records can be quickly created with the harvesting facility. This feature, especially when using the shortcuts, is very fast, and allows a more flexible approach when records can be created without even entering the CORC site (password access is sufficient). Automated features also have potential, even if the resulting records currently require careful editing by cataloguers.

CORC libraries have the option of leaving records in the OCLC database or exporting them into their own local catalogue or Web site. There are benefits in relying on the services of an external agency for record storage, but libraries may prefer an integrated approach with records loaded into the catalogue, which then becomes more of a one-stop shop.

Subject gateways are another way of searchers finding quality selected Web resources.

How does CORC compare to subject gateways such as AVEL (2000) and AGRIGATE (2000)? They are similar to the extent that the resources in them are selected by experts, they are cooperative ventures, and mostly rely on metadata. They are different in that subject gateways are dedicated to specific subjects or subject areas, while CORC spans the full range of subjects. Gateways are free while CORC is a fee-paying service with different rates for full members and partial users. The CORC community will thus be limited – mainly to institutional members and their communities.

Because CORC is using new standards, it necessarily suffers in terms of needing to constantly revise and update the system. CORC and other services using metadata have to co ntend with the problems associated with using evolving schemas. Dublin Core is as yet unfinished and may never be, with various communities still selecting qualifiers specific to their needs. CORC is still defining its own metadata standards for use in its system.

What will be the optimum format for web-based resources? Will MARC persist or will Dublin Core or some other format prevail? Dillon in his paper for the LC Bicentennial Cataloging Conference (2000) discusses what he sees as the options – use or adapt MARC/AACR2, start afresh, and use or adapt Dublin Core. He concludes with the suggestion of a MARC version of Dublin Core to suit the aims of libraries. He argues that it would be easier to adapt Dublin Core than to adapt AACR2/MARC. McCallum of the Library of Congress’s MARC Office (2000) says that because of the growing number of electronic resources and the potential to unite the resources and cataloguing data, we need to experiment with "differentiation and selection of resources for levels of control; reevaluation of descriptive content requirements for cataloging, and the exchange record format structure" (2000: 2). Another suggestion proposed by Weinheimer of Princeton University Library (2000:a) is that cataloguers and Web site creators should globally cooperate to catalogue Web material. The basic notion that Web creators could very well describe their pages/sites and that cataloguers are best placed to decide points of access sounds reasonable. However such global cooperation between these two groups seems unlikely to get off the ground. Such proposals, however, do need to be considered, and who knows which ones will eventually bear fruit and develop into realistic projects and ventures.

The very nature of Web resources is causing cataloguers and others to rethink their standards. For example, monographic Web sites begin to look a lot like serials when they keep changing. Should they be treated like loose-leaf publications or as a brand new category? Such issues are being considered by cataloguing committees globally. The broken Web links will be reported by CORC to members and users, but individual libraries will still need to fix the links themselves. This is of course a concern for libraries and others whether they use CORC or not. How do we keep the links up-to-date in a satisfactory way? Many CORC issues are, therefore, also those which need to be faced in the broader community.

While CORC and its enthusiasts are riding high on the wave of the new system, there is another view being put forward that search engines will improve to such an extent that no one will need to catalogue Web resources. Baruth (2000) argues that third generation search engines will make projects like CORC obsolete. She believes that improved filter algorithms will bring users better results, making it the best option for searchers. Only time will tell whether CORC and its equivalents will pr evail or be necessary. Weinheimer (2000:b) in an editorial letter regarding an article by Arms (2000) which speculated regarding how automation might take over many of the skilled tasks currently carried out by librarians (e.g. the search engine, Google), argues that good catalogue records, created once for all to use, and accompanied by a good reference structure, are in fact cost effective. He also queries whether "popular" sites are necessarily quality sites. Librarian-selected resources, as in the CORC catalogue, do provide for some quality control, and users have the benefit of access points and cross references, which they now enjoy in catalogues.

Most librarians agree on one thing – that new services are essential if libraries are to remain relevant. As Dempsey (1999: 5-6) states "Libraries now operate in a shared network space which brings together users, services providers and resources in new combinations and balances Unlike some other player s, libraries have a dual role. They manage their own collections and make them available, but they also have a role in guiding their users to resources and services out of their direct control." "The hybrid library cannot be a mere collocation of services, a listing on a web page. Where is the added value the library provides? The value of the library is that it saves users’ time, that it releases the value of the resources it manages, that it effectively brings together users and resources over time" (1999: 4).

CORC is, therefore, one of many projects attempting to bring users and resources together. It is also one of the more significant initiatives in terms of its scope, its size and international involvement. OCLC because of its substantial research base and firm footing in the US has the resources to test new technologies and to try to be a world leader in the area of cataloguing Web resources.

The UNSW Library decided to participate in the CORC Project and to continue cataloguing on the new system as a way of investigating the option of making the catalogue the one-stop shop for selected Web-based resources. The project was undertaken in the spirit of experimentation and in the knowledge that additional expenditure in terms of CORC fees and staff time would be involved with little return in the short term. While the pathfinder creation tools proved of no value at this time to us, they will be reviewed in the future.

The cataloguing system with automated features including a harvesting option provides for quick initial record creation. While cataloguers must still edit and upgrade records for inclusion in our MARC-based system, alternative ways of cataloguing (e.g. in the Dublin Core format) may be a future option for certain material.

The project encourages participants to focus on the role of the catalogue and to consider how one can best provide access to online resources. The issue does remain, however, of whether search engines will eventually improve to such an extent that cataloguing Web resources will become unnecessary in the longer term. Will Baruth (2000) and like-minded thinkers prove to be correct? At this stage it appears that libraries cannot really afford to wait and see. Cooperative ventures provide an opportunity to investigate options for assisting our users to find the full range of resources they are seeking. The project and system described here is just one of those that the UNSW Library will be participating in. At the time of writing the Library remains open-minded regarding CORC, and is interested in continuing some involvement to explore the issues as they arise.

AGRIGATE: An Agriculture Information Gateway for Australian Researchers. (2000). [Online]. Available: http://www.agrigate.edu.au/html/index.html [2000, October 27]

Arms, William Y. (2000). How Effectively Can Computers be Used for the Skilled Tasks of Professional Librarianship? D-Lib Magazine, July/August 2000, vol. 6, no. 7/8. [Online]. Available: http://sunsite.anu.edu.au/mirrors/dlib/dlib/july00/arms/07arms.html [2000, October 27]

Australian Digital Theses Program. (2000). [Online]. Available: http://adt.caul.edu.au/ [2000, October 26]

Australasian Virtual Engineering Library (AVEL). (2000). [Online]. Available: http://avel.edu.au/ [2000, October 26]

Baruth, Barbara. (2000). Is your Catalog Big Enough to Handle the Web? The Case Against OPAC Integration of Online Resources. American Libraries, August 2000. vol. 31, no. 7.

Cooperative Online Resource Catalog (CORC). [Online]. Available: http://www.oclc.org/oclc/corc/index.htm [2000, October 26]

Dempsey, Lorcan. (1999). The Library, the Catalogue, the Broker: Brokering Access to Information in the Hybrid Library. [Online]. Available: http://www.ukoln.ac.uk/dlis/models/publications/landscape/ [2000, October 26]

Dillon, Martin. (2000). Metadata for Web Resources: How Metadata Works on the Web. [Online]. Available: http://lcweb.loc.gov/catdir/bibcontrol/dillon_paper.html [2000, October 26]

Hickey, Thomas B., Childress, Eric, & Watson, Bradley C. (1999). The Genesis and Development of CORC as an OCLC Office of Research Project. OCLC Newsletter, May/June 1999.

Irvine, Duncan T. D. (1999). CORC Project Participants Hold First Meeting. OCLC Newsletter May/June 1999.

McCallum, Sally. (2000). Extending MARC for Bibliographic Control in the Web Environment: Challenges and Alternatives. [Online]. Available: http://lcweb.loc.gov/catdir/bibcontrol/mccallum.html [2000, October 26]

MetaChem. (2000). [Online]. Available: http://metachem.ch.adfa.edu.au/ [2000, October 26]

University of New South Wales Library. (2000). CORC Project. [Online]. Available: http://www.library.unsw.edu.au/~eirg/corc/corc.html [2000, October 26]

University of New South Wales Library. (1999). Metadata Template. [Online]. Available: http://www.library.unsw.edu.au/~eirg/metatemp.html [2000, October 30]

Weinheimer, James L. (2000:a). How to Keep the Practice of Librarianship Relevant in the Age of the Internet. [Online]. Available: http://www.princeton.edu/~jamesw/metadata [2000, October 26]

Weinheimer, James L. (2000:b) [Letter to the Editor] D-Lib Magazine, September 2000, vol.6, no. 9 [Online]. Available: http://sunsite.anu.edu.au/mirrors/dlib/dlib/september00/09letters.html [2000, October 27]