Home // ICSEA 2017, The Twelfth International Conference on Software Engineering Advances // View article

From Language-Independent Requirements to Code Based on a Semantic Analysis

Authors:
Mariem Mefteh
Nadia Bouassida
Hanêne Ben-Abdallah

Keywords: Natural Language Processing; Semantics; Requirements; Naturalistic programming; Syntactic/Semantic grammar

Abstract:
This paper presents a new approach, which allows building Java codes from language-independent requirements. In other terms, it does not require any manual transformation of the requirements into the syntax of a specific programming language. To handle these challenges, our approach relies on a set of English-based mapping rules to generate a semantic representation of the input requirements. This semantic representation is used to produce the source code via existing code generators, such as Pegasus_f. Indeed, our approach extracts the Pegasus code from the semantic representation. This code is refined by eliminating the redundancy among the code elements’ names thanks to the Term Frequency/Inverse Document Frequency (TF/IDF) method and the Density-based spatial clustering of applications with noise (DBSCAN) algorithm. Finally, Pegasus_f transforms automatically the resulting Pegasus code to Java. The proposed approach is implemented through the Code Recovery tool (CodeRec-tool), which accepts the English and the French languages in its actual version. The simplicity and the usefulness of our approach have been evaluated using measurements and based on experts’ feedback.

Pages: 145 to 156

Copyright: Copyright (c) IARIA, 2017

Publication date: October 8, 2017

Published in: conference

ISSN: 2308-4235

ISBN: 978-1-61208-590-6

Location: Athens, Greece

Dates: from October 8, 2017 to October 12, 2017