Extracting Visions from Textual Requirements Documents

Miriam SayãoCarlos A. Prolo

Software requirements are frequently written as a document in natural language. For large systems it is difficult to index them so that we can extract information concerning a particular point of view of interest. Indeed they frequently come with no interesting indexing system. In this paper we investigate the use of information retrieval techniques based on natural language processing to produce indexing information for requirements documents written in natural language. We focus on two aspects: one is how to retrieve the collection of requirements related to a particular user-oriented view of interest. The other is how to produce a suitable indexing system for the requirements document, which an expert software engineer would see as a general and natural way of accessing the document for purposes ranging from understanding the system to maintenance, to requirements verification and validation.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: