THE ABERYSTWYTH INDEX LANGUAGES TEST
| Date | 01 January 1973 |
| Published date | 01 January 1973 |
| DOI | https://doi.org/10.1108/eb026547 |
| Pages | 1-35 |
| Author | E. MICHAEL KEEN |
| Subject Matter | Information & knowledge management,Library & information science |
THE
Journal of Documentation
VOLUME 29 NUMBER 1 MARCH 1973
THE ABERYSTWYTH INDEX LANGUAGES TEST
E. MICHAEL KEEN
College
of
Librarianship
Wales,
Aberystwyth
Reports a laboratory comparison of the effectiveness and efficiency of five
index languages in the subject area of library and information science;
three post-co-ordinate languages, Compressed Term, Uncontrolled, and
Hierarchically Structured, and two pre-co-ordinate ones, Hierarchically
Structured and Relational Indexing. Eight test comparisons were made,
and factors studied were index language specificity and linkage, indexing
specificity and exhaustivity, method of co-ordination, the precision devices
of partitioning and relational operators, and the provision of context in the
search file. Full details of the test and retrieval results are presented.
WHAT WAS UNDER TEST
THE PURPOSE OF this research test was to compare under laboratory
conditions the effectiveness and efficiency of a number of different kinds of
index language as used in subject retrieval systems. The subject-field was
library and information science, and the five languages tested were:
Compressed Term Index Language, post-co-ordinate (CT)
Uncontrolled Index Language, post-co-ordinate (UL)
Hierarchically Structured Index Language, post-co-ordinate (HS)
Hierarchically Structured Index Language, pre-co-ordinate (Pre-HS)
Relational Indexing Index Language, pre-co-ordinate (Pre-RI)
This choice of languages covered those in frequent operational use and
one or two others strongly advocated by their proponents, but for research
purposes a feature of greater importance
was
that quite wide extremes were
represented in the degree of vocabulary control and amount of effort re-
quired in the operation of each language. Various measures of vocabulary
1
JOURNAL
OF DOCUMENTATION Vol.
29,
no. 1
Index
language
Compressed Term (CT)
Uncontrolled (UL)
Hierarchically Structured,
post-co-ordinate (HS)
Hierarchically Structured,
pre-co-ordinate (Pre-HS)
Relational Indexing
(Pre-RI)
TABLE
I:
Index
languages under test
Origin
Adapted from
a
term list
compiled
by
Aslib's Research
Department
Vocabulary of words manu-
ally extracted from
the
abstracts of the documents
Adapted from A
Classifica-
tion
of
Library
Science,
Classi-
fication Research Group, 1965
Devised by J. Farradane
Method
of use
Post-co-ordinate
Post-co-ordinate
Post-co-ordinate, with
terms unsynthesized,
but
retaining notation of the
faceted classification
Pre-co-ordinate, with terms
synthesized into classmarks
according
to
prescribed
citation order
Pre-co-ordinate, with terms
synthesized into analets using
a two-dimensional display
Primary method of
indicating cross-references
Thesaural relationships,
mostly
RT
None
Hierarchical display
Hierarchical display
Hierarchical display
Special features
Small vocabulary of less
than 300 terms
Almost no vocabulary
control
Faceted classification
Faceted classification
Nine relational operators
to specify relationships
between subject elements
2
March
1973
INDEX LANGUAGES
control will be presented, but Table 1 provides a general comparison of the
five languages as to their origin, method of
use,
method of showing cross-
references, and other special features. The objective of the test was to make
a series
of comparisons of different sets of the languages, usually in order to
isolate just one variable at a time. Eight such comparisons were made, and
the purpose of each comparison together with the languages involved are
given
in
Table
2.
It may be seen that variants of some of the languages were
required, labelled
CT1,
CT2, and
so
on,
so
that other factors such
as
changes
in indexing specificity or indexing exhaustivity could be investigated.
Special controlled search comparisons were needed to isolate the precision
devices of partitioning and relational operators, as well as the provision of
context in the search
file.
In describing the different languages and the com-
parison tests the abbreviations used in Table 2 will be employed: details of
the main language variants will appear in Table
3.
It may be noted here that
the full project report1 gives an account of additional experiments not re-
ported here, including the identification of search variables, a comparison
of search
rules,
the design of a search strategy experiment, and desiderata for
retrieval effectiveness measures.
TABLE
2.
Test comparisons made
Factors
in
comparison
(1) Post-co-ordinate index
languages of measured
specificity and linkage
(2) Indexing specificity
(3) Indexing exhaustivity
(4) Method of co-ordination
(5) Post-co-ordinate index
languages including variations
in indexing specificity and
exhaustivity
(6) Precision device of
partitioning
(7) Precision device of relational
operators
(8) Provision of context in the
search file
Index
languages
CT1,
UL1,
HS2
UL1,
UL2
CT1,
CT2
HS1,
Pre-HS
CT1,
CT2,
UL1,
UL2, HS2
HS1,
Pre-HS (Partitioning)
Pre-HS (Partitioning),
Pre-RI (Relational operators)
HS1,
HS1 (Context)
Results
Tables 8, 9 &
Figure 3
Tables 10 & 11
Tables
10
& 11
Tables 10 & 11
Table 12
Table 13
Table 13
Table 13
3
To continue reading
Request your trial