SRC: UNVERIFIED ASSUMPTIONS

The speaker explained that the stimulus for the SRC proposal came partially from the UNISIST study report in which the working group responsible for subject specification had drawn attention to the need for better tools for the control and conversion of information retrieval languages, such as UDC and to the need for further study and experiments. The working group had already considered a study undertaken at their request by a team from the Aslib Research Department under Mr B. C. Vickery in which, however, only the term content of different schemes, mainly in the English language and including UDC, were compared. When this investigation was carried out in 1969, only twenty of the expected one hundred parts of the English Full edition were available, whereas now there are eighty. Even so, UDC was less defective (including over 80 per cent) in content of terms than the other major general schemes of classification studied, i.e. BC, CC and DDC. The study also included a comparison between the classificatory relations used in some major thesauri, i.e. the EJC/TEST, NLM/MeSH and EURATOM thesaurus, with those used in the general classifications already mentioned. As far as UDC was concerned, this did not compare as favourably as more recent studies by Kara, Öhman and Ölivecrona, and Stueart, and Wellisch. This was partly because the Aslib study did not take into account the well‐known synthetic characteristics of UDC, i.e. the use of colon combinations and of common and special auxiliary subdivisions. He maintained that FID should not have embarked upon a programme of radical revision based upon the proposals for an SRC (see Aslib Proc., Vol. 24, no. 4, April 1972, p. 222–5) because: