Application of type theory and linear logic for Programming Languages

点击翻译按钮获取中文摘要

Type THeory is provides an ideal formal framework for program verification and systematic program development utililzing the logical mechanism. On the other hand, linear logic provides the framework for expressing computational resources and concurrent computation. Therefore, it seems very promising to combine these two theories in order to design a powerful and new programming language and its program development tools. For this purpose, we have investigated the theoretical foundations for this combined formal framework, linear type language.In particular, we established verious important semantics theories for the linear type language. For example, phase semantics is extended to a dynamic phase semantics in which cut-elimination and normalization proofs can be performed. Phase semantics is also extended to higher order linear type languages. Girard introduced Light Linear Logic, which is considered an impotant sybsystem of linear logic since it characterized the polynomial time computation in a purely logical way. We applied our phase semantics theory to Light Linear Logic and established phase semantic characterization of the important concepts used in Light Linear Logic. The coherence semantics has also been very much improved. For example, Coherence Banach Space Semantics, which is a promising denotational semantics for the linear type language, was developped when the French team leader, J-Y.Girard, visited our group in Japan. Various important computation models for the linear type language have been proposed in our project research. In particular, the reduction paradigm and the proof-search paradigm were studied from the programming paradigm and the proof- search paradigm were studied from the programming languages point of view.