Technical Tips Ellen C. Rappaport Albany Law School erapp@mail.als.edu

Some Tales of ALA 1999 Midwinter and Annual Meetings:
I picked up my copy of the newly-revised Anglo-American Cataloguing Rules at ALA’s Midwinter meeting, in January. Its cover is a pale, diffident green, unlike the bold hunter green of the 1988 edition. After attending this year’s ALA meetings, I suspected that the tentative color may represent the fleeting nature of this edition.

A major topic, in January and in June was: What is a serial, and how should we describe (catalog) it? Over the past several years, the CONSER Task Force on AACR has discussed the nature of seriality. Jean Hirons of the Library of Congress has played a leading role in this discussion, and is the chief author of the Task Force’s April 1999 paper “Revising AACR2 to Accommodate Seriality: Report to the Joint Steering Committee on the Revision of AACR” [available at www.nlc-bnc.ca/jsc/ser-rep0.html]. She spoke at Midwinter, at numerous meetings at ALA’s June conference, and at AALL. In her paper, seriality (the serial nature of library material, or bibliographic resources) is defined thus:

“Seriality refers to the fact that certain types of bibliographic resources are issued over time, regardless of whether they are intended to be complete or have no predetermined conclusion, and thus, the bibliographic data can change over time.”

This is broader than the present AACR2 definition of a serial: “A publication in any medium issued in successive parts bearing numeric or chronological designations and intended to be continued indefinitely.”

If implemented, the new concept of seriality will change the way we catalog a lot of the material in law libraries. The big difference is in recognizing the serial nature of loose-leaf titles, sets with supplementation, and Web sites. If we change the way we catalog them, our online systems may treat them differently.

The paper presents this model of seriality:

Static resources (no seriality): Static items are complete when they are first issued (or published), whether in a single part (a one-volume novel), or in multiple parts (a two-disk compact disk recording). This category includes books, e-texts (such as e-books), maps, sound recordings, etc.

Successively-issued resources: Information is added to successively-issued items in a succession of discrete parts. This category includes serials, series and multi-part items that are not complete when they are first issued, such as a large treatise published in 8 volumes over five years. It also includes monographic titles (themselves static) which are updated by a succession of supplements.

Integrating resources: Information is added to integrating resources in a succession of changes which do not remain discrete. This category includes loose-leafs, Web sites (made up of many changing Web pages), and changing databases.

For years, the cataloging rules considered loose-leafs as monographs (they sit in a single binder), while we struggled with their added pages and changing authors, titles, and publishers. Law librarians have always known that loose-leafs had seriality. Like loose-leafs, Web sites may not have successive issues, they may not have incrementing numbering or chronology, and when they change title, they sometimes drop the old title completely. These loose-leaf characteristics are troubling to catalogers accustomed to normal periodicals and serials, but they’re old hat to law library catalogers whose loose-leafs have been integrating for decades! Now as the rest of the world’s catalogers deal with Web sites, the concept of integrating resources may solve our problems with loose-leafs.

Hirons’ paper proposes to develop cataloging rules that allow us to describe library material which is static, which is successively-issued, and which is integrating. It recommends changes to AACR2 which will accommodate all these forms. For integrating resources, it proposes to begin with Adele Hallam’s Cataloging Rules for the Description of Loose-leaf Publications, adding rules as needed for electronic resources. This is good news for law catalogers who have long relied on Hallam’s work.

We may get to be less picky about what is a title change. In fact, it is recommended that we think less about what constitutes a title change, more about what requires a new bibliographic record. Title changes may not necessarily generate a new bib record. By June, it looked as though print periodicals and series will continue to receive successive entry cataloging (a separate bibliographic record for each title change). But integrating entities such as loose-leafs, Web sites and additive databases may be cataloged on one bib record under their latest title and latest publisher, with earlier titles and publishers added. Electronic journals (e-journals) which change title may be cataloged by successive entry (separate records) if the earlier title still appears on the earlier issues. But those e-journals which display only the latest title (even for the old issues which were originally issued under the earlier title) may be cataloged all on one bib record under the latest title.

Another paper which was discussed at ALA in June was MARBI Discussion Paper no. 114 “Seriality and MARC 21” [available at lcweb.loc.gov/marc/marbi/dp/dp114.html]. MARC 21 is the name for the newly-combined U.S. and Canadian MARC formats, which provide computer labels, or tags, for cataloging records. If AACR2 is changed, MARC coding must change too.

That pale green cover of AACR2 may reflect an effort to reorganize the cataloging rules. The Task Force is urging a three-part book: Part I could be description, no longer organized by physical format, but by area of description, more like most MARC-tagging manuals. Since physical formats overlap these days, they are no longer an unambiguous way to organize the rule book — the cataloger can no longer be sure which chapter to start with. (Is it a book or is it electronic? Is it a sound recording or a serial?) Part II could again be concerned with access points. Part III could deal with relationships to other works: linking entries and uniform titles. Some of the thinking underlying this proposal is in Tom Delsey’s complex paper “The Logical Structure of the Anglo-American Cataloging Rules drafted for The Joint Steering Committee for the Revision of AACR” (August 1998), available at http://www.nlc-bnc.ca/jsc/aacrdel.htm.

Comments on the Hirons paper can be submitted to the Joint Steering Committee through September — instructions for commenting are on the JSC Web site, along with the paper. The JSC will meet in the fall to continue these discussions.

Godfrey Rust’s presentation at the annual meeting (during Midwinter) of the National Information Standards Organization (NISO) gave us all a reason to pay attention to metadata. He connected many of the standards being developed for identifying works and users.

Metadata is a general term for structured data which describes other data. A cataloging record is our best-known example of metadata: it describes a particular book; it can be linked to one or more unique identifying numbers. The term “metadata” is now being used mostly for machine-generated or author-generated data which describes electronic resources, such as Web sites. Ideally, metadata is carried within the electronic resource, so that the resource comes with all the data needed to find it and to track its use.

To allow us to find (or otherwise deal with) a book or other object, it must have some kind of unique identifier (perhaps a number), and a description that goes with that identifier so we know — with certainty — what is being identified. In the library and publishing worlds, we already have standards that identify and describe the things we purchase, collect and lend: ISBN and ISSN, UPC codes, AACR and USMARC, Dewey and LC call numbers, and many others. To some degree, these traditional forms of metadata allow unique identification and description. But they were designed for specific uses by libraries, booksellers and publishers, to describe relatively simple resources which were stored in a physical location. Some require significant intellectual work and time to create. And, while these standards have rules and structures, they don’t fully relate to one another.

Rust stated that we need a new kind of metadata. It must be useful for many purposes. It should be generated when the creation is first distributed and used over and over again (“do it once, do it right”). It should be capable of dealing with all kinds of material. It must deal with complex aspects of that material: a book is only taken off a shelf, but a digital object may be copied, excerpted, reformatted and distributed in a new medium, purchased or licensed with limitations.

Rust put it this way:

* PEOPLE make STUFF. (Authors write books, singers record songs, etc.)

* PEOPLE use STUFF. (Readers read books, fans listen to concerts, etc.)

* PEOPLE make deals about STUFF. (Readers buy books and subscribe to magazines, fans purchase recordings, readers borrow books and articles through Interlibrary Loan, readers photocopy articles, restaurants play recordings of songs, publishers reprint photographs, people download text, data, video clips, sound clips, etc.)

Rust suggested that now that articles, sound recordings, video clips and databases are readily available electronically, in our libraries and on the Web, we need far tighter control over who may use our resources than a library card that allows you to borrow a book. He maintained that authors, publishers, booksellers, providers of databases, and libraries share the need to identify material (STUFF), its users (PEOPLE) and the terms on which it is used (DEALS). Rust spoke of creators and of creations. He defined a right as “authority originating in law for a Party to do or to authorize another Party to do a defined act to a Creation.” He spoke of rights-owners and of rights administration. He reminded us that libraries, although we may not charge for use, still have agreements by which we purchase and lend material.

He insisted that, with electronic availability of creations, rights must be controlled automatically. There will soon be far too many transactions to handle it manually. We will need standard identifiers tied to descriptions (metadata). The advantage of the new kind of metadata (as opposed to traditional cataloging) is that it can be built into the creation, and perhaps can be generated automatically.

Rust named several dozen organizations now working on schemes for metadata to identify and describe book components, creations of all kinds, sound recordings, musical works, audiovisuals, etc. This is an exciting moment, he said. Because many standards are moving toward the same goal, there’s hope of interoperability — of compatibility which will allow us to use the various standards to generate the metadata we need to allow the use of material electronically. “Metadata is the lifeblood of e-commerce,” declared Rust.

Slides of his presentation are available on NISO’s Web site (http://www.niso.org/pdfs/rust.pdf). His related paper “Metadata: The Right Approach : An Integrated Model for Descriptive and Rights Metadata in E-commerce” is available at http://www.dlib.org/dlib/july98/rust/07rust.html.