On Nov. 17-19, 1997 I attended the second of OCLC's seminars entitled "Knowledge
Access Management: Tools and Concepts for Next-Generation Catalogers" (http://www.oclc.org/institute/seminar2.htm). Although the title sounds general, the seminar focused on the intersection of cataloging, metadata, and the Internet. There were 22 of us, many from academic libraries, both large and small, but also some from a network, a large public library, a consulting firm, a special library, and a library science faculty.

There were 3 broad themes to the seminar: libraries and the Internet, issues in the cataloging of Internet resources, and the role of different kinds of "metadata" in the online environment. I have tried to highlight below some of the major points made for each theme.

LIBRARIES AND THE INTERNET

This session was an argument that the phenomenal growth of the World Wide Web in the last year or so has brought with it many of the scholarly resources that libraries traditionally collect, and thus libraries will have to "collect" and catalog these new Web-based versions of these resources in the future.

• The WWW grew from 1993-1996 to include 30,000 sites; from 1996 to the present we have gone to 1 million sites. Growth in telecommunications capacity remains a major driving force in today's world, leading to Scott McNealy's remark that Internet years are just like dog years, seven to every one.
• Why all this WWW growth? The multimedia capabilities provide a better educational experience.
• The downside of all this WWW growth: the Web's too disorganized to search effectively (just when our major need is for organizing information, library schools are bypassing cataloging in the curriculum!). The library field has created some wonderful organizing tools like the MARC record, classification schemes, authority control.
• Good cataloging is hard to do: differentiating each publication from all the others is not easy and will only get harder. Nevertheless, we must begin to mainstream Internet resources in our cataloging workflow.

This theme was ended by a prediction that any OPAC vendor who does not have a Web catalog interface by the end of this year will be out of business.

ISSUES IN THE CATALOGING OF INTERNET RESOURCES

These sessions were conducted by Ann Sandberg-Fox who has been involved in both national and international discussions of bibliographic description of electronic resources.

We went through every cataloging rule and MARC field used in cataloging Internet resources discussing difficulties and possible solutions all along the way. Basic bibliographic concepts are hard to apply to Internet resources (an example she used that illustrates many of the issues discussed below is http://www.oc.ca.gov/):

• Monograph vs. Serial: We have hung a great deal on our traditional division of material into monograph and serial, a distinction that doesn't translate well to Internet resources. As things now stand, Web sites are cataloged as "monographs".

• Title: Even the title of a website isn't always clear from the layout and the problem is compounded by the fact that pages may be displayed differently by different Web browsers. There is already some reliance being put on the title as recorded in the <title> area of the html document source code, even though it doesn't always correspond exactly to anything as viewed on the Web page!

• Edition: Is "new and improved" on a home page to be taken seriously as an edition statement, the way it would be if printed in a book? Probably not. One wonders if edition statements as such will disappear for Internet resources, with a date being the only fixed point in identifying "versions of websites."

• Date: …Speaking of dates, which dates? Often "last update" is the only date that appears. Some copyright dates are appearing and these would probably take precedence for identifying "editions." In describing Websites we may need to use the "serial" note: "Description based on [last update date]."

• General principles: Some of the general principles that are emerging in discussions both of Internet resources and the future of the cataloging rules are:

• All Internet resources are considered "published."
• The international community favors the term "electronic resource" over our current "computer file" format term
• More emphasis on "identifying" resources over exact transcription of their text
• A move away from format-based cataloging: users like information gathered together on one record.
• A new definition of "seriality"

• Bibliographic record of the future: these may be records that are permanently "unfinished." Some identifying features would be permanent but other parts that we might want in the latest form (e.g., contents) would have a pointer to the website to search, retrieve and display the up-to-the-minute information from the website, on-the-fly.

• Library of Congress Internet projects

• Table of contents (TOC): this project follows well on the "record of the future" idea described above and is connected to LC's Electronic CIP (E-CIP) program. A method was developed to take TOC information from the E-CIP manuscript, wrap minimal HTML coding around it, save it to the WWW server, and add the 856 field to the catalog record. The catalog record has as its "contents note" a URL to the contents on the Web. In another project, LC is automatically adding the first subject heading to the TOC site in a keyword field. This means that contents words can be found by Web search engines and link back to the catalog record which could then serve as a source for finding related items. David Williamson from LC, who was an attendee at the seminar, reported on several TOC projects involving interaction between catalog records and Web sites (sample LC record with Web URL: OCLC#32822869).

• BEOnline (http://lcweb.loc.gov/rr/business/beonline/beoabout.html): this is a pilot project to provide access to business and economic resources from the Internet. The plan is to have it serve as both a model and a catalyst for developing approaches to identifying, selecting, and providing bibliographic access to remote electronic sources on the Web. This should begin to yield LC copy for Internet resources.

• OCLC and the Internet

• Intercat database (http://purl.org/net/intercat): This is a searchable database of more than 20,000 catalog records for Internet resources. It was begun by OCLC as part of a grant project sponsored by the Dept. Of Education, but continues to be maintained. There is also an Intercat listserv (http://www.oclc.org/oclc/forms/listserv.htm) for continuing discussion.

• PURLs (http://purl.oclc.org/): PURLs are "persistent URL"s. OCLC is offering to maintain an "authority file" of URLs with the PURL serving as a cross-reference or alias for the real URL, which is, hopefully, maintained by the person who registered for the PURL. PURLs can be assigned automatically at the Website.

METADATA

Metadata is a term being used to refer to any data about other data. Catalog records are currently libraries' most familiar form of metadata but this is an area of great interest, excitement and experimentation and other forms of metadata are emerging to serve different needs.

• Why all this interest in metadata?

First, the impetus is not coming from libraries, who have always recognized the importance of metadata by creating catalog records. The impetus is coming from the WWW community: the Web browsers, the search engines, and the Internet service providers.

We all know what searches on the Web are like! A recent and very good metadata survey article by Warwick Cathro of the National Library of Australia (http://www.nla.gov.au/nla/staffpaper/cathro3.html) described a Web search on the acronym IETF (Internet Engineering Task Force). Admittedly, this is an Internet body and will have a lot of hits, but, still, it's not as general as searching for a common subject word like "environment." Nevertheless, the "IETF" search retrieved 896,354 matches! Cathro went on to explain that metadata would allow search engines to target searches onto words or phrases that identify their correct role, e.g., "green" as a personal name vs. "green" as a subject. This sounds elementary and it is, but it is still more than many search engines can do.

• DUBLIN CORE (DC): a new form for metadata (http://purl.oclc.org/metadata/dublin_core/)

In March, 1995, the first Metadata Workshop was convened at OCLC, and included researchers and professionals from librarianship, computer science, text encoding, and related areas, to come up with a set of core metadata elements to describe networked resources. The result was called the Dublin Core, "Dublin" because of OCLC's location in Dublin, Ohio, and "Core" because this was meant to have the minimal number of elements required to describe an electronic resource adequately.

Since then, there have been four more Dublin Core conferences with these results:

• The DC consists of 15 elements, all of which try to have a commonly understood meaning : (for more details: http://purl.oclc.org/metadata/dublin_core_elements)
  • title
  • author or creator
  • subject and keywords
  • description
  • publisher
  • other contributor
  • date
  • resource type
  • format
  • resource identifier
  • source
  • language
  • relation
  • coverage
  • rights management
• It has been determined that the DC will be adequate to describe both documents and images.

• Many packages of metadata should be able to exist side-by-side. This is referred to as the "Warwick framework" which is "a conceptual model for a container architecture for metadata packages of various types. The organizing principle that emerged from this workshop was that there will be many packages of metadata, independently developed and maintained for different purposes by various communities, and that this modularization will allow for coherent evolution of different components of the metadata landscape under a single metadata architecture." (from 4^th DC Metadata Workshop report: http://www.dlib.org/dlib/june97/metadata/06weibel.html)

• Many experimental projects are using the DC (for details: http://purl.oclc.org/metadata/dublin_core/projects.html)

Now that the DC has become standardized, various "crosswalks" are being developed between DC and other types of metadata, especially the MARC record (http://lcweb.loc.gov/marc/dccross.html). However, there are still many questions whose answers are speculative:

• Where will DC records come from?

In the long run, from people with a vested interest in better access to Web items, e.g., Website creators, database creators, catalogers, etc. There is also software already available for creating DC records for existing Websites (for fun with this try DC Dot: http://www.ukoln.ac.uk/metadata/dcdot/).

• How soon will we see DC-compliant browsers?

6 months, as a conservative estimate. For a recent report: http://www.zdnet.com/pcweek/news/1103/03rdf.html

• Other forms of metadata in use in libraries

• TEI header: The Text Encoding Initiative (TEI) is an international project that guides the SGML-encoding of electronic texts for scholarly research. At the top of the file for each encoded text is a TEI header that describes the text and its original source, and documents the encoding history (http://www.uic.edu/orgs/tei/). Our own project Documenting the American South includes a TEI header for each text, and this is then used as the basis for the MARC record when each text is cataloger-a good example of coexisting forms of metadata.

• GILS (Government Information Locator Service): identifies and describes information resources throughout the US Federal Government (http://www.access.gpo.gov/su%5Fdocs/gils/gils.html).

• EAD (Enhanced Archival Description): used by libraries and museums for archival finding aids (http://sunsite.berkeley.edu/FindingAids/EAD).

CONCLUSION

The MARC record, as old and imperfect as it is to us, is viewed by others as "very rich metadata," with all of its coding and authority control. No one expects the DC to replace the MARC record although it is likely that catalog records and other types of metadata will merge eventually. Right now, there is room for all kinds of metadata. We may find ourselves using several kinds in one catalog: the MARC record for the most important and scholarly resources, with, perhaps, the DC becoming a new minimal level format.

The seminar opened with some memorable quotes from library history, this one from Cutter in 1904: "I cannot help thinking that the golden age of cataloging is over, and the difficulties and discussions which have furnished an innocent pleasure to so many will interest them no more." If Cutter were here to see us at this new stage of catalog development I think he would be pleased.

Celine Noel

Catalog Dept.

Davis Library

1-29-98