Abstract not found

We define a specification formalism (formally, an institution) which provides a notion of dynamic type (the type which is associated to a term by a particular evaluation) and late binding (the fact that the function version to be invoked in a function application depends on the dynamic type of one or more arguments). Hence, it constitutes a natural formal framework for modeling objectoriented and other dynamically-typed languages and a basis for adding to them a specification level. In this respect, the main novelty is the capability of writing axioms related to a given type which are not required to hold for subtypes, hence can be "overridden" in further refinements, thus lifting at the specification level the possibility of reusing code which is offered by the object-oriented approach.

this technicality

This paper describes the logic of the RAISE Specification Language, RSL. It explains the particular logic chosen for RAISE, and motivates this choice as suitable for a wide spectrum language to be used for designs as well as initial specifications, and supporting imperative and concurrent specifications as well as applicative sequential ones. It also describes the logical definition of RSL, its axiomatic semantics, as well as the proof system for carrying out proofs.

Abstract not found

The temporal logic institution

this paper. Study of PSPs has long tended in the direction of PDTs [1, 2, 3, 5]. One of the problems is that, while the former continued the tradition of working with classes axiomatized by (possibly conditional) equations, the latter require a precise grasp on individual algebras (which, for modeling purposes, can be identified with programs): a program P taking as a parameter another program X cannot change X -- X functions in the context of P , that is in P [X ], in the same way as it would in isolation. This intuition of "preserving actual parameter" has been identified as one of the semantic requirements on PSP in form of the persistency requirement on the functors from # Email: yngvel@ii.uib.no + Email: michal@ii.uib.no (1), e.g., [3, 15, 2]. However, in the purely equational context, there was hardly any syntactic counterpart of this semantic requirement. Thus, no syntactic/logical means were available for reasoning about correctness of such implementations

Abstract not found

We propose a formal specification framework for functional aspects of services. We define services as operations which are specified by means of pre- and postconditions, for the specification of which we use extensions of description logic. The (extensions of) description logic and the specification framework itself are defined as institutions. This gives the framework a uniformity of definition and a solid algebraic and logical foundation. The framework can be used for the specification of service requests and service providers. Given a signature morphism from request to provider, we define when a service request is matched by a service provider, which can be used in service discovery. We provide a model-theoretic definition of matching and show that matching can be characterized by a semantic entailment relation which is formulated over a particular standard description logic. Thus proofs of matching can be reduced to standard reasoning in description logic for which one can use description logic reasoners.

There are a number of reasons why software construction is an inherently hard process to master. Specification plays a central role here; therefore, better means of specification improve productivity. One way of achieving this may be the use of formal specification languages, which have the advantage of being unambiguous. This thesis proposes and motivates a new formal specification language called AFSL (Almost Formal Specification Language or Another Formal Specification Language). AFSL is not a full-grown tool that can be used productively in the software industry right away. Rather, its development has been a quest for creative ideas that may enhance the applicability of formal specification in the future. This introduction explains what formal specification is and argues why it is useful, discusses the historical background of AFSL, and gives an overview of the rest of the thesis.