Currently, the EDC extends the vocabulaire file with all metadata information concerning the database tables. When the database contains many tables, and only few are used in Corticon the vocabulaire is cluttered with unnecessary information. As a result, performance on vocabulaire actions is slow and compilation is slow. When the metainformation is stored in a seperate file, the vocabulaire remains compact and clean. Another option would be to have a clean option that removes unused data, but that seems a suboptimal solution.