"The Catalog As Portal To the Internet"
by Sarah E. Thomas

commentary by Brian E. C. Schottlaender

Final version

Ms. Thomas' good and thoughtful paper, and its assorted "modest" proposals, challenge us to ask ourselves eight questions-some of which are embodied in her closing recommendations, others of which are not. I shall, in this commentary, respond briefly to each in turn.

Q: Should libraries create and manage a mechanism to support access to Internet resources?

In fact, libraries already do manage many such mechanisms. We have all no doubt heard of, and perhaps even used, Cornell University's Gateway, the University of California San Diego's Sage, and the University of Wisconsin's Scout-to name a few. The problem, as I know Ms. Thomas knows, is that libraries have not created "a [single] mechanism." We have created several. As a consequence, not only must we manage several, but our clients must navigate several.

Q: Are the tools used to build the catalog appropriate for description of Web resources?

The tools used to build the catalog will work (well enough) to describe Web resources, but it may not always be appropriate (or desirable) to use (all of) them all of the time.

Q: Should the catalog serve as the portal to the Web?
A: NO.

It can't, it shouldn't, and it doesn't need to. In fact, catalogs and portals are "metadata constellations" which, when integrated-as, for the most part, they presently are not-make up, along with other such constellations, the "universe of access." To ask catalogs to serve as portals to the Web is asking too much of them, just as asking portals to serve as catalogs of "the non-Web" is asking too much of them. I do not believe that I disagree with Ms. Thomas in taking this position inasmuch as her own argument speaks of "reinterpreting" the catalog, and of imagining a "new information medium" that is a hybrid combining some of the best features of the catalog and the portal. She is, thus, no longer talking about the catalog.

Q: Should libraries "decisively reduce the amount of time we devote to the cataloging of books in order to reallocate the time of our bibliographic control experts to provide access to other resources, especially Internet resources. . . ?"
A: NO.

Ms. Thomas' "InfoGlut" slide describes an international publishing environment in which book production is averaging one million volumes published annually. In that sort of environment, libraries will not be in a position to "decisively reduce" the amount of time devoted to book cataloging, at least not until such time as it has been demonstrated that:

Further, Ms. Thomas' recommendation unnecessarily and undesirably dichotomizes between books and all other information resources. Libraries were uncomfortable with the "access vs. ownership" dichotomy; Michael Gorman has suggested that the "cataloging vs. metadata" dichotomy is a false one1; this one is no better.

Finally-and ironically-Internet resources lend themselves to automated processing ("cataloging") in ways that [printed] "books" do not. And yet, we've barely begun to explore how best to take advantage of these automated processing capabilities. Better, at this point, we should continue to explore that avenue than go down that of reducing our cataloging commitment to those materials which do not lend themselves to such processing.

Q: Should libraries investigate and implement a combination of the following:

a. using the PCC core bibliographic record;
b. using Dublin core or a modification thereof;
c. accepting copy with little or no modification from other cataloging agencies, including vendors;
d. working with publishers, authors, and software developers to encode publications in a standard way that permits the generation of metadata from digital objects through the use of software programs;
e. increasing collaborative efforts nationally and globally so that publications are cataloged according to mutually acceptable standards in a timely fashion and once only?

It is highly desirable that libraries pursue any and all of these strategies, although not to "reduce the time spent cataloging books" specifically, but, rather, to maximize the time spent cataloging generally. Of these, (a), (c), and (e) in combination have the most promise, while long-term investment in (d) may or may not yield a long-term dividend.

Q: Should libraries increase the functionality of their catalogs/portals by:

a. increasing the scope and coverage of materials;
b. ensuring timely access to publications;
c. increasing the level of access from citation to full-text or increasing degrees of granularity;
d. incorporating features such as reference linking, recommended titles (others who liked this title also liked:), relevance ranking, customization, and personalization that "make portals so captivating?"

(a) and (b) above are non-controversial because libraries should always have provided timely access to materials of sufficient scope and coverage to meet the needs of their clientele. Unfortunately-and this is no doubt part of Ms. Thomas' point-"libraries should always have" does not mean "libraries have always." (c) above is more debatable because what level of access is appropriate when and to whom is itself (or are themselves) debatable.

Finally, (d) above is especially debatable because if one polled the 100+ attendees at the LC Bicentennial Conference on what features they think "make portals so captivating," one would probably not only get 100+ different answers, but half the attendees would probably disagree with the other half. One person's "captivating" is another's annoying, erroneous (cf. the amusing discussion in Thomas Mann's paper of the "recommended title" feature found on a number of portal sites), or presumptuous. Ms. Thomas' slide describing corporate portals as "competing for the eyeballs of their [i.e., the corporations'] employees" is symptomatic. Would that portals sought to compete for the minds of those using them, rather than our eyes!

Q: Should libraries not "go it alone," but instead:

a. collaborate with other libraries in a coordinated plan for the acquisition, creation of metadata, access, and preservation of materials available through portals;
b. define a clear path from the local library portal to the larger scholars portal;
c. partner with developers of portals and search engines to share expertise in a constructive way, drawing on the best each has to contribute to goal of effective access to information?

Collaboration and partnering both facilitate standards development and implementation and reduce unwanted redundancy. Libraries have long been fairly good at cooperating with each other. As noted by Priscilla Caplan in her paper, however, we've not been terribly good at collaborating with those outside our own community. Just as, for example, the librarians who developed the EAD DTD did so in concert with DynaWeb's software developers, so too will it behoove libraries in general to work with Intel and the like to pursue the development and refinement of Internet discovery mechanisms. It is worth noting that Ms. Thomas' "Manage the Knowledge of Thousands" slide depicts not a confident young librarian striding into the future, but, rather, a Lotus employee doing so!

Q: Should libraries bring their "light" out from under the "bushel basket" by:

a. advertising the features of the discovery database, a hybrid combining some of the best features of the catalog and the portal;
b. quantifying the value of the laborsaving features of the portal/catalog for its clientele;
c. seeking new revenue (from partner portals?) in order to expand their scope and accomplishments;
d. conducting and publishing research that documents improved results through use of the catalog (saves time, finds more appropriate materials; titles found are accessible, etc.)?

Researching, quantifying, and publicizing ("advertising") the features of the "discovery database" and its time and scope implications are good ideas. Whether the discovery database is a physical construct-as Ms. Thomas' characterizing it as "a hybrid combining some of the best features of the catalog and the portal" makes it sound-or a logical construct which amalgamates into a single view (though, as suggested in Thomas Mann's paper, perhaps not a "seamless" view) the range of relevant resources offered up by each remains to be seen.

Notes 1.)Michael Gorman. "Metadata or Cataloging?: A False Choice." Journal of Internet Cataloging 2(1): 5-22.

Library of Congress
December 19, 2000
Library of Congress Help Desk