Engineering Terminology -a Case for a Linguistically- Informed Terminology Database

Terminology databases of specialist domains contain a wealth of lexical, semantic and pragmatic data associated with each of the terms stored. However, until recently, the amount of syntactic and morphological data associated with each term is either nonexistent or entered as ad-hoc grammatical data, like compound nouns, adjectival nouns, etc. The recently completed EC-sponsored TRANSTERM project has demonstrated that through the use of a linguistically well-grounded model and an accompanying toolbox, it is possible to enrich and re-use terminology resources in a number of applications ranging from terminology management to information retrieval. The toolbox comprises a corpus manager and a morphosyntactic analyser that increases the overall performance of the toolbox.