ISR11
Scientific Report No. ISR-11 Information Storage and Retrieval
Design Criteria for Automatic Information Systems
chapter
M. E. Lesk
G. Salton
Harvard University
Gerard Salton
Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government.
V-2
b) the intellectual aids to be used as part of the manual analysis
and indexing procedure, including dictionaries, thesauruses, and
hierarchical subject arrangements are best prepared and maintained
by c[OCRerr]iuittees of experts in the subject areas under consideration;
c) the users of the service, being unaware of system restrictions
and operations, should not submit search requests directly to
the system but imist work through human intermediaries who analyze
the query statement and prepare suitable search formulations for
introduction into the program.
A system organization based on these principles leads to a service in
which oniy the search operations themselves are mechanized (that is, the
comparisons between analyzed information items and analyzed search requests),
but most other operations are carried out semimanually or manually. It
also results in an information system which suffers from so many built-in
weaknesses that adequate service to the users cannot ever be expected.
The first weakness is the well-known scarcity and increasing unavaila-
bility of subject experts who are willing and able to perform a manual
content analysis of the documents and search requests. This simple
fact results in a continuing crisis atmosphere in existing nonconventional
search systems, a situation which may be expected to grow more severe as
time progresses. The second weakness is the inadequacy of the presently
available dictionaries and authority lists which are used to control the
assigrmLent of subject identifiers to the stored information. These dictio-
naries are often produced as a result of so many compromises among various
expert committees, that the final product reflects no consistent point of
view, and is difficult to utilize effectively. The third weakness is the
absence of meaningful user interaction with the system, so that individual