Abstract

Constraint diagrams are a visual notation designed for use by software engineers to formally specify information systems. In this paper we formalize a fragment of the constraint diagram language. A set of reasoning rules are defined and we prove that this set is both sound and complete. Given constraint diagrams D1 and D2 such that D2 is a semantic consequence of D1, to prove completeness we construct a proof of D2 from D1. A decision procedure can be extracted from this proof construction process and it follows that the system is decidable.

Received 25 January 2005.