We present an interactive query formulation technique which enables exploitation not only of structural properties of data but also of semantic constraints as posed by the contents of data. The technique aims at the formulation of a semantically consistent or meaningful query by the end-user without any previous knowledge of syntax formalisms and data model semantics. This has been achieved by end-user guidance in that an inference engine suggests semantically rich query terms for further consideration by the end-user. The set of suggested terms at each interaction stage comply with the already considered query terms with respect to structure and contents based semantics. Assignment or selection of operational terms are also allowed, if operational semantics comply with the semantics of data. The interactive query formulation component has been implemented in Java and runs on the client side of a client/server based query answering system architecture. |