Modular semantics for object-oriented models

  • Authors:

  • Ali Hamie;John Howse;Stuart Kent

  • Affiliations:

  • Division of Computing, University of Brighton, Brighton, UK;Division of Computing, University of Brighton, Brighton, UK;Division of Computing, University of Brighton, Brighton, UK

  • Venue:
  • 3FACS'98 Proceedings of the 3rd BCS-FACS conference on Northern Formal Methods
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes a formal, modular approach to the semantics of object-oriented models. These models are expressed in the UML notation augmented with the Object Constraint language (OCL), which provides a precise textual language for expressing assertions: invariant constraints and operation specifications. The approach to semantics we adopt could easily be applied to other, similar notation sets. An important aspect of our approach is to treat every component of a model, down to the level of individual diagram elements, as distinct semantic entities which, here, are theories in Larch. The semantics of a model or part of a model is then the composition of the semantic entities corresponding to the individual model elements. This leads to a highly modular approach, allowing reuse of generic semantic entities, resulting in an elegant, transparent semantics for individual models. The fine-grained modularity promises to support the extraction and manipulation of different "views" of a model. The composition of model elements supported by the semantics promises to underpin the development of systems from reusable components. Larch is used as it supports theory composition, ensures that the semantics is relatively accessible (essentially theories of FOPL), is well-defined, and is close to technologies most likely to provide automated tool support (including a toolset to support it directly).