Abstract. Software requirements specification is a critical activity of the software process, as errors at this stage inevitably lead to problems later on in system design and implementation. The requirements are written in natural language, with the potential for ambiguity, contradiction or misunderstanding, or simply an inability of developers to deal with a large amount of information. This paper proposes a methodology for the natural language processing of textual descriptions of the requirements of an unlimited natural language and their automatic mapping to the object-oriented analysis model.

No abstract