Synthesizing ASMs from natural language requirements
Vincenzo Gervasi · 2001
State Machines (ASMs) have been successfully used in a number of studies (see [2] for a survey) to specify and model various software-intensive systems. Among the properties that make use of ASMs desirable for this purpose, the ease with which an ASM can be inspected by the customer or nal user of the system plays an important role. In fact, the clear operational semantics of ASMs makes it easy even for non specialists to understand the behavior described by an ASM model. In this paper, we report on the dual approach, that is, synthesizing an ASM model from a natural language description of a system. We show how relevant information can be extracted from informal requirements in natural language, and how the information obtained can be used to produce an ASM model on a level of abstraction comparable to that of the original requirements. The initial model can then be refined in a controlled way to add further details, and possibly obtain the executable code for the system in the end. ...