The completion semantics for finite failure
Christopher John Hogger · 1990
Abstract The construction of a consequence•-oriented semantics for the finite failure extension is more problematic than for pure Horn-clause programs and queries. In the latter case we have a particularly simple connection between logical meaning and operational meaning: As soon as we include Jtiil calls in our language no such connection applies--- a program containing fail is not a construct of classical logic and so is not amenable to the notion of classical logical consequence. Nevertheless, a variety of analogous connections have been suggested. The best-known of these is based upon the so-called completion semantics, which relies upon two ideas: