XFM: extreme formal method for capturing formal specification into abstract models

  • Authors:

  • David Berner;Syed Suhaib;Sandeep Kumar Shukla;Jean-Pierre Talpin

  • Affiliations:

  • INRIA, project Espresso, IRISA, Campus de Beaulieu, Rennes Cedex, France;Department of Electrical and Computer Engineering, Virginia Tech, Blacksburg, Virginia;Department of Electrical and Computer Engineering, Virginia Tech, Blacksburg, Virginia;INRIA, project Espresso, IRISA, Campus de Beaulieu, Rennes Cedex, France

  • Venue:
  • Formal methods and models for system design
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this chapter we introduce an agile formal method (named XFM) based on extreme programming concepts to construct abstract models from a natural language specification of a complex system. Building formal models for verification purposes is being used in the industry for two different usage modes: (i) Descriptive Formal Models (DFM) are used to capture an implementation into an abstract model to submit to analysis by model checking tools, (ii) Prescriptive Formal Models (PFM) are used to capture natural language specifications into a formal model to analyze consistency of the specification and also as a reference model to compare a DFM against it. We propose XFM as a methodology to incrementally build a correct PFM from a natural language specification. We illustrate the benefits of the proposed methodology with the help of two examples: a control intensive traffic light controller, and the DLX pipeline. Our experiments show that this methodology not only constructs abstract models in sufficiently shorter time than the time taken in constructing ad hoc abstract models from implementation or specification, but also provides models that are constructively correct, closer to the intended specification and better structured.