The thesaurus is managed by the Oracle RDBMS, and is accessed from a C++ program using embedded SQL, over a TCP/IP network. The database structure and access methods are based upon those used for the CILKS project described earlier in section 1.3.1.
