Comments on Tom Delsey's Paper, "The Library Catalog in the Networked Environment" for the Library of Congress Bicentennial Symposium "Bibliographic Control in the New Millennium", Washington, DC, November 15, 2000

Jennifer Trant, Executive Director,
Art Museum Image Consortium and Partner, Archives & Museum Informatics

Final version

I was reticent about accepting an invitation to comment at a Library of Congress conference on the future of cataloging in the new millennium, for I am neither a library cataloguer nor an information scientist. My knowledge of bibliographic tradition parallels my knowledge as a Canadian, of the American electoral system: alternating glimmers of understanding and total bafflement.

But when I considered the topic, I began to think that I might have something to bring to the discussion. I am, for example, an experienced cataloguer (though the things that I learned to describe are works of art). I've been involved in consortial efforts to define the nature of a "catalogue description" (the Categories for the Description of Works of Art), and I now direct a consortium (AMICO - the Art Museum Image Consortium) that is building what might be considered a union catalog. Many of the issues that have been addressed in these discussions parallel those of the library community - but they've taken place in a parallel universe.

I'd like to highlight some of these issues within the context of the Tom Delsey's subtly expressed repositioning of the library catalog in the networked environment. He challenges us to rethink the nature of the catalog and what it describes. To fully understand the implications of this repositioning, we need to consider how the catalog is used, and when it is called upon in the research process.

The Nature of the Catalogue

First of all, I'd like to thank Tom for shifting the focus of the discussion from "bibliographic control" to the nature and purpose of the catalogue itself. The phrase "Bibliographic control" conjures up narrow connotations of the physical management of a distinct kind of object. The "networked catalog" shifts the our attention to a the role of description and access in the nexus between the collection and the user. But we have to be careful about the tempting inversion "cataloging the Web", for it may be the same kind of malapropism as "MARC cataloging". Just as librarians don't catalog with MARC, they don't describe the Web, but information resources that are available on the Web.

Conceiving of the catalog as an interface to networked information requires re-examining its content and its structure.

Many museums are now facing this question, as they consider the implications of putting 'collections online'. They are struggling to re-purpose collections management systems into public access systems. With both the catalog and the "object" delivered digitally, the boundaries between the catalog and the information resource it describes have become blurred. What was once considered a tool for managing a physical collection is becoming a means for presenting knowledge about that collection. With this change must come a parallel adjustment in the content and nature of the catalog itself. We are learning to wrap the individual descriptions of discrete objects with the context helps users make meaning. The data elements that enabled us to answer the questions 'What is it?" and "Where have I put it" don't fully satisfy the need to interpret the nature of a digital object, and help the researcher understand what a museum object means, and how it relates to other things.

What does the Catalog Describe?

Traditionally, the library catalog reflects a series of selection and acquisition decisions, based upon a collections development policy. Tom's paper hints that the networked library catalog may contain additional information. And he points us to the possibility that all of this information might not have to be supplied by cataloguers. Data from abstracting and indexing services may provide a level of granularity not achievable given the economics of traditional cataloging. Analytics virtually integrated into the catalog from other sources would allow the researcher to gain access to the unit of information appropriate for her task (for she'd like to find the article not the journal issue). Links to online texts enable the delivery of the resource itself.

Another challenge to the traditional conception of the bibliographic control, however, is that the resources likely to be used in a networked information space may or may not be "bibliographic". Formally published writings are now integrated with drawn or digitally photographed images, recorded sounds, reconstructed models, mathematical simulations in a fluid digital space and these new genres require a different methods and structures for their description.

These new genres raise new issues: much of the information required to adequately document such networked information resources is extrinsic to the resource itself. Even if a digital object could in some way technically 'self-describe' through a declaration of embedded metadata, much intellectual description becomes a matter of assertion. The catalog record begins to represent opinion, rather than fact. The catalog itself, becomes a publication -- its contents a compilation distinguished by their selection, arrangement, authority, authenticity and interpretation. This is certainly the direction the AMICO Library (http://www.amico.org/) is moving. Interestingly these characteristics are shared by the kinds of catalog that one often encounters in the art world: the Exhibition Catalog, the Permanent Collection Catalog and the Catalog Raisonnée all embody scholarship and opinion as much as they represent 'fact'. As document genres they sit on the boundary between metadata and data itself.

How is the catalog used?

It was helpful to be reminded in Tom's paper of the Functional Requirement s for Bibliographic Records: to find, identify, select and obtain. These processes seem to have their online equivalents in the realm of information discovery and retrieval. Much discussion of metadata in the Web environment has focused on this first step in the research process, finding relevant resources. (Cross-domain resource discovery was one of the motivators of the Dublin Core initiative.)

But information discovery isn't an end in itself. The act of obtaining information doesn't answer the researchers' question. Users of networked information resources have come to expect seamless support of their entire research process. The catalog and the information resource blur when the interface to both is the web browser. The digital library catalogue becomes the means for the delivery of dynamic digital content, requiring us to reassess the functional requirements of the library catalog. We're asking those catalog records to do much more than they used to! Tom hinted at this when he spoke about access management. Here, we could look to the archival community, who have long administered restrictions on access to collections, and to museums, whose relationships with contemporary artists offer another model.

The intersection between the catalog and the research process

Repositioning the catalog requires a model of when and where it is used. In a paper presented at a UK Office of Library and Information Networking Conference in 1998 David Bearman and I explored the inter-relationships between metadata requirements and the Humanities research process (paper online at http://www.archimuse.com/papers/ukoln98paper/index.html). We identified five broad phases in this process: Discovery, Retrieval, Collation, Analysis and Representation. Building on the ideas expressed in the Warwick Framework we wondered what kinds of metadata would be required at which point, in this iterative process, and how it might be supplied by a well designed library catalogue working in a networked information system.

Discovery and Retrieval are phases we are familiar with. Finding (Discovery) takes place in the public space of the networked library catalog. The researcher identifies resources of interest, finds the available copies that are nearest, or most convenient, or most suitable. During retrieval, this content is moved from the library space to the users space, to enable further use. Already, in supporting digital retrieval we may be adding requirements for technical elements to our metadata, those required by the user to judge which of many formats might be most appropriate.

Librarians have not concerned themselves with what users do, and often philosophically denied any knowledge of this area of activity. But we need to consider use if we are to support it. Use can be broken down into a series of individual functions with specific characteristics, each of which require or generate metadata: collation (integrating new resources into existing ones), analysis (deriving, creating or assigning meaning) and re-presentation ( the publication or re-distribution of new knowledge. This process is cyclical; the act of re-presentation creates a new resource to be discovered. If metadata is managed throughout this process, then the description of the new resource is much easier.

Different kinds of metadata will have a role to play in each of these phases. Much of this is metadata that is likely to find its way into a the catalog of a distributed digital library. Further research and discussion about the nature of that catalog and the way that users interact with and use digital information resources and their descriptions is critical to the creation catalogs with utility. The value that the library profession could add to these new kinds of catalogs may not be in locally created catalog records -- other authors have drawn our attention to initiatives producing metadata along with networked information resources and the bar-code scanning cat distributed with Wired magazine is now touted as a way to catalog books (see http://www.wirednews.com/news/gizmos/0,1452,39139,00.html)

Seeing the role of the library catalog as that of a mediator and provider of access to networked information, rather than as a management tool for a repository of books requires a more active management of resources and relationships. Architectural solutions than enable the incremental integration of disparate and distributed resources are key to enabling us to "fast track" the building of catalogs of networked information (pouring concrete in the foundation before all the details are designed is not unusual in large architectural projects). What is key is up-front exploration of the knowledge structures of the disciplines we serve, and how do they intersect.'.

Throughout our discussions of the future of the library catalog we need to remember that users don't search the catalog in order to find a catalog record. They aren't even really looking for a book -- they are looking for information resources that help them answer their questions and accomplish their tasks. The challenge for the digital library catalog is to provide the right information about useful resources, at a time and in an environment that supports user processes. Looking outward is key to repositioning the library catalog within the networked information environment.


Library of Congress
January 02, 2001
Library of Congress Help Desk