The MULTILEX linguistic architecture is meant to supply the backbone to the dictionary design. It defines the blocks of features and rules the type of relations they can bear. The relations restricted to the "containers" allow the flexibility of the content - the features -, the selection of which is a function of the needs of the languages involved, of the designers' projects and of the available data and expertise.
The MULTILEX linguistic architecture combines the lexicographic approach with the multi-dimentional facilities the electronic media offers. The organization of the data is allowed to provide the logical and linguistic ground for building user interfaces which would give access to linguistic facts in a way which is natural to lexicographers for the tasks they perform dayly. It also permits to safely and easily define sub-models derived from the initial linguistic architecture for extracting consistent sub-sets of the information from the main resource to a specific application lexicon.