School of Information Management & Systems. Spr 1997. M. Buckland. Infosys 101: Information Systems.
Description: Entities, Attributes, Values, and Context.

Bibliographic access can be viewed as "fitting descriptions". How well does the searcher's description of what is wanted fit the "system's" descriptions provided in the bibliography, catalog, other information system? Bibliographic descriptions should anticipate how searchers' questions will be expressed. Searchers need to pose their questions in terms of the descriptions used in the system. A description of an entity (object, book, person, concept, etc.) is composed of combinations of attributes ("color of eyes") and values ("brown"). What the values signify depends on the context because (i) the values are meaningful only in relation to some undertood standard and because (ii) to assign values is to make descriptive and/or evaluative statements which is a language activity, which in turn depends on the cultural context for meaning. Which attributes will be most useful to provide in any given (purposive) retrieval system? How can attributes and values be expressed in a mutually intelligible and mutually convenient ways? How much description is worthwhile given limited resources? Some features of attributes are:

1. "Brute facts": Parts of what is being described, e.g. title on title-page, original abstract, text,...

2. "Tidiness", e.g. Can fit a fixed standard format (e.g. year of publication); use standardized codes (e.g. language: ENG SPA JAP). Examples: Book ID# (e.g. International Standard Book Number ISBN; International Standard Serial Number ISSN); Catalog record ID# (e.g. Library of Congress Card Number LCCN. Each system (MELVYL, OCLC, RLIN) ordinarily has its own ID numbers for each record. Can it be expressed as YES or NO (Is it a biography?).

3. "Messiness", e.g.:
-- of variable length (e.g. title, author's name)
-- absent, present once, or present many times (e.g. authorship: Anonymous? Many authors?)
-- complex structure (e.g. complicated names)
-- open-ended (e.g. explanatory notes)
-- ambiguous (e.g. alternative names and subject descriptions)

4. Varied, e.g. (a) Copy of (part of) the object (e.g. picture of it, text); (b) Description of the object (physical dimensions, provenance); (c) Description of what the object signifies (what the text is "about"); (d) Description how one object is related to another (e.g. continues, comments on, contradicts).

Distinguish between (i) the description of an object, mostly a selective summary of details derived from the object, but can include description derived from other sources; and (ii) the "access points" (indexing terms, subject headings) used to avoid having to search through all of all of the descriptions. For a book on Albania look for the subject heading ALBANIA, then look at the descriptions of the books listed under ALBANIA. What about alternative headings, e.g. variant names? A standardized list of headings is called a "Thesaurus" or "Authority File". Standardizing headings is called "Authority control" or "Vocabulary control".

Conventions and standardized methods for making and handling (a) descriptions (attributes, values) and (b) headings are needed. In libraries the Anglo-American Cataloging Rules 2nd ed (AACR2) is a highly stylized, agreed standard for describing documents. AACR2 incorporates the International Standard Book Description (ISBD), a standard for the visual presentation of bibliographic descriptions by imposing special punctuation on the description. MAchine-Readable Cataloging (MARC) denotes a growing family of standards for formatting and for communication between computers of records used for various purposes in information retrieval. The MARC display of Benny and his Penny implements all three. (Revised 2/14/97)