In this project, we study the problem of processing a software requirements specification expressed in natural language. We observe and classify linguistic phenomena in a corpus of specification examples. Also, we explore systems presented in related literature and identify their most useful characteristics. From these, we design and implement a method for interpreting a specification, expressed in a very simple yet informative format, and producing a labeled transition system (LTS). Our strategy combines an interactive approach and ad hoc decision heuristics. Representations are enriched by processing semantic phenomena. Several lines are proposed for its further extension, including an annotation model in the educational context.

This project was the foundation of an university-incubated start-up, and my participation during the time I worked in this company led to my Masters dissertation.


If you are interested in more of this, please read my M.Sc. thesis (in Spanish), for the technical details of the system architecture, and the instantiation and analysis of semantic, linguistic and knowledge-related phenomena.