School of Information Management & Systems. Spring 2004.
245
Organization of Information in Collections.
Michael Buckland.
Assignment 7: Make a Small Thesaurus.
11A: Sample due March 8;
11B: Thesaurus due March 15.
A thesaurus generally has single terms (or phrases) suitable for a
searcher to combine in Boolean
searches (aka "post-coordinate indexing"). Create a small thesaurus
for vocabulary control. This could be for an
attribute in the collection you are to describe.
Past examples.
1. Read
Morville, Peter.
Building
a Synonymous Search Index
and his
How
Do You Build a Thesaurus?. (Following the links is
optional. One of them, a link to examples
is meant to lead to the American Society for
Indexers "Thesauri online site" which is now at
http://www.asindexing.org/site/thesonet.shtml
visited in Assig. 7.)
2. Make explicit assumptions about
the scope and purpose:
The situation in which it is to be used;
the purpose of its use; and
the user group to be served.
3. Use the following two formats as applicable:
Entry for Preferred Terms: e.g.
HAIR [Preferred term]
Scope Note [Definition, explanation, scope -- unless obvious]: Includes fur of animals
Use For: TRESSES Non-preferred term
Broader [preferred] Term: HEAD
Narrower [preferred] Term: EYELASHES
Related [preferred] Terms: SCALP, WIG
(In this exercise, to reduce effort, use italics for BT, NT, and RT terms for which you are not
establishing entries.)
Entry for Non-preferred Terms: e.g.
TRESSES Use HAIR
4. Deliverables: See Taylor 276-283 (1st ed, p. 157) for models.
11A: Draft Sample: List of assumptions,
examples of likely
terms, sample Preferred Term with SN, BT, NT, etc.
11B: Thesaurus. Minimum of 15 Preferred Terms,
with list of assumptions,
sources used (Where did your terms come from?),
alphabetical list of Preferred Terms and Non-preferred Terms, with their
relationships as in #3,
hierarchical listing of Preferred Terms (only) with decimal notation,
and a diagram of relationships between Preferred Terms, either Tree Structure(s)
or Arrowgraph(s).
Notes:
Only in a very simple case does a thesaurus form a single hierarchy.
Analysis in terms of facets, each likely to form a separate
hierarchy, commonly helps.
In any thesaurus expected to last, one should have a History Note for each
term indicating when it came into use, its source, and when it was retired.
The School does not yet support any specialized thesaurus construction software
and it is not needed at this small scale. However, anyone wishing to download
and use thesaurus software at their own responsibility on to their own
machines could consider "TheW32," available at as freeware from
http://publish.uwo.ca/~craven/freeware.htm