A Statically Typed Logic Context Query Language With Parametric Polymorphism and Subtyping

Abstract

The objective of this thesis is programming language support for context-sensitive program adaptations. Driven by the requirements for context-aware adaptation languages, a statically typed Object-oriented logic Context Query Language (OCQL) was developed, which is suitable for integration with adaptation languages based on the Java type system. <br /> The ambient information considered in context-aware applications often originates from several, potentially distributed sources. OCQL employs the Semantic Web-language RDF Schema to structure and combine distributed context information. <br /> OCQL offers parametric polymorphism, subtyping, and a fixed set of meta-predicates. Its type system is based on mode analysis and a subset of Java Generics. For this reason a mode-inference approach for normal logic programs that considers variable aliasing and sharing was extended to cover all-solution predicates. <br /> OCQL is complemented by a service-oriented context-management infrastructure that supports the integration of OCQL with runtime adaptation approaches. The applicability of the language and its infrastructure were demonstrated with the context-aware aspect language CSLogicAJ. CSLogicAJ aspects encapsulate context-aware behavior and define in which contextual situation and program execution state the behavior is woven into the running program. <br /> The thesis concludes with a case study analyzing how runtime adaptation of mobile applications can be supported by pure object-, service- and context-aware aspect-orientation. Our study has shown that CSLogicAJ can improve the modularization of context-aware applications and reduce anticipation of runtime adaptations when compared to other approaches.