The Habanera Lexical Knowledge Base Management System

Habanera is a multipurpose multilingual lexical knowledge base that is developed at CRL to be used as a central repository of multilingual lexical data. The knowledge base contains a set of dictionaries and relations between entries, within a dictionary (e.g., synonymy) as well as between entries of different dictionaries (e.g., translation). The format of monolingual lexical entries is left relatively free: the toplevel follows the TEI structure, and lower-levels are dictionary and language dependent. An entry is encoded using a feature structure representation. Feature structures are typed thus enabling a formal definition of the dictionary schema; type definitions are also used by the database typechecking tools.