A NATURAL-LANGUAGE APPROACH TO FLEXIBLE MODEL QUERYING

We describe an interface to a crop model that uses techniques from nat ural language interpretation work. A graphical user interface (GUI) an d a conventional command language each has its own limitations in term s of expressive power. Our restricted natural language interface (NLI) has some of the advantages of real natural Language, such as familiar ity of medium and tolerance of the user's manner of expression, togeth er with some of the advantages of the GUI. Behind the interface itself , inferences are made with respect to the model, including query valid ity checking and disambiguation. in this paper we consider relevant pr ior research, before describing the project aims and details of the in terface. We conclude with a discussion of the strengths and weaknesses of the approach taken, and how it could the improved.