ldeas and proposals for the development of retrieval languages in OMIKK

Abstract

Automation plans in the National Technical Information Centre and Ubrary (Országos Műszaki Információs Központ és Könyvtár) and the establishment of its databases require development of retrieval languages. Free-text search and automatic indexing are not sufficient to cover all kind of information retrieval tasks: retrieval languages having controlled vocabulary are also necessary. Various factors have influence on these languages, multiple purpose indexing of collections being one reason for difficulties. An attempt to meet all needs by a single complex retrieval language seems to be a failure, one needs more mutually harmonized languages, that is, a mixture of natural language based systems and classification schemes. The library needs ongoing use of UDC for the planned COM-catalogue and a natural language based system for the online catalogue. The file of abstracts requires use of a set of special thesauri, in addition to free text search and UDC-based search factlities. As far as their contents is concerned, all these thesauri should be coordinated by a common roof thesaurus called „Classaurus" (in Hungarian: Osztaurusz); as to structure and methodology of thesaurus establishment, they are to be coordinated by the common use of a well-proved Hungarian software package for thesaurus construction and maintenance (TSPR). The work is to be done step by step. All individual development phases would result in well-defined products (simple lists of keywords - lists of descriptors with equivalence relations - roof thesaurus - special thesauri) available in and applicable to various system-elements as retrieval languages.