From Code to Models

  • Authors:
  • Gerard J. Holzmann

  • Affiliations:
  • -

  • Venue:
  • ACSD '01 Proceedings of the Second International Conference on Application of Concurrency to System Design
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

One of the corner stones of formal methods is the notion that abstraction enables analysis. By the construction of an abstract model we can trade implementation detail for analytical power. The intent of a model is to preserve selected characteristics of real-world artifact, while suppressing others. Unfortunately, practitioners are less likely to use a modeling tool if it cannot handle real-world artifacts in their native format. The requirement to build a model to enable analysis is often seen as a verdict to design a system twice: once in a verification language and once in an implementation language. Because the implementation phase cannot be skipped, verification is often sacrificed. In this paper we will consider a way to avoid this problem by automating the extraction of verification models from implementation level code. The user now provides only model extraction rules, or abstractions, rather than full-scale models.