School of Information Management & Systems. Spring 2004.
245
Organization of Information in Collections.
Michael Buckland.
SUBJECT ACCESS IN A CATALOG.
1. People don't always agree what a document is about.
A matter of point of view, perspective.
2. Relying on the words in a document (title, text, etc.)
has limitations.
Very common words are usually
ignored as a hindrance: A, THE, OF, FROM..,... "Stop words".
3. You need language for description
- and language is culturally based. Use of "natural
language" (ordinary use of words) results in inconsistency
since there are unlimited different ways
to describe something. "Unnatural language" can
provide consistency, by
(a) "Controlled", limited, standardized, and stylized
use of words (List of subject headings,
thesaurus, controlled vocabulary); or
(b) Artificial notation to denote and arrange subjects,
e.g. a library or other classification using numerals.
Communities have their cultures and so vary in specialized
terminology, concepts, views, and
values, so description needs to be situational:
for some group, for some purpose.
3. Combinations of topics,
e.g. "The physics of music." The cataloger/indexer isolates the
different aspects ("facets") of the subject.
How do you find a particular combination?
(a) The indexer can assign a descriptor for each
separate aspect of the subject matter. The
searcher examines the list of books associated with each facet,
then compares the lists to see
which books occurs on all of these lists, i.e. has
that combination of descriptors. Check the list of
books on "Physics" and the list of books on "Music"
and identify the books on both lists.
("Post-coordinate" searching).
This "Boolean searching" is typical of online
bibliographic databases.
(b) Alternatively, the cataloger can define and label
each combination before anyone searches
("Pre-coordinate indexing"): (i) General books on
Agriculture in one list; (ii) General books on Mali
in another list; but (iii) Books on "Agriculture in Mali"
(a combination) only in a separate third list.
For manual searches in card catalogs and printed indexes
conventions are needed govern
arrangement (File under "Agriculture--Mali" or
"Mali--Agriculture"?). This has carried over into
"eXact SUbject" searches in online library catalogs.
N.B. Pre- and post-coordinate systems should be logically equivalent.
(c) More powerful "second generation" online library
catalogs, e.g. MELVYL, PATHFINDER, use a curious
combination. The indexing data available is precoordinate
subject cataloging from the Library of
Congress Subject Headings list. So these headings can
be searched as composed (as a "phrase").
They are also decomposed into individual words
indexed individually so that the words can be
searched separately or in any combination as specified by
the searcher, i.e. post-coordinate
searching, using a "keyword" search (e.g. Subject keyword).
Thus a post-coordinate ("subject keyword")
searching capability in provided as well as exact
pre-coordinate "phrase".
4. Other clues, e.g. title words,
language, date, and authorship are often useful, e.g. Find
records for a few books on a subject using a search on title
keywords, then use a full display a
to reveal the subject headings.
5. Library catalogs use the Library of
Congress Subject Headings, in big red volumes
usually kept
near online terminals, which makes elaborate use of
subheadings to denote subtopics, form (e.g.
"dictionaries"), place, time, etc. There are also
cross-references, e.g.
USE [= use some other heading] and its reciprocal
USE FOR - for synonyms.
Dowsing USE Divining-rod; Divining Rod UF Dowsing.
Usually relates synonyms.
NT = "See also the Narrower Term" and its reciprocal
BT = "See also the Broader Term": Dogs NT Spaniels;
Spaniels BT Dogs.
RT = "See also the related term": Birds RT Ornithology.
SA = "See also": Dog breeds SA names of specific breeds,
e.g. bloodhound, collies,...
Ability testing SA subdivision Ability testing under subjects,
e.g. Dentists--Ability testing.