A Step Towards Ambiguity Less Natural Language Software Requirements Specifications

In modern software engineering practice, the ability to specify ambiguity less software requirements in a natural language (NL) in a seamless way is highly valuable and desirable. Though, the software requirements are typically captured in natural languages (NL) such as English, there is a very high probability that more than half NL requirements can be ambiguous. For example, Mich identified that approx. 72% of the NL requirements are potentially ambiguous. A primary reason of such ambiguous NL requirements is syntactic and semantic ambiguities in a natural language such as English. A problem with ambiguous NL requirements is that a software engineer can miss-interpret requirements and can generate an erroneous and absurd software model. In this paper, we aim to address this challenge by presenting a novel approach that is based a semantically controlled NL representation for software requirements. To generate a semantically controlled NL representation, we propose the use of Semantic of Business Vocabulary and Rules (SBVR) standard. We solve a case study to bear out that a SBVR based controlled representation can not only help in generating accurate and consistent software models but can also simplify the machine processing of requirements. The results show that our approach can be helpful in generating the accurate and consistent software models from NL software requirements. A Java implementation of the used approach is also presented a proof of concept that is also available as an Eclipse plugin.