An Intelligent Analytics Approach to Minimize Complexity in Ambiguous Software Requirements

Abstract

An inconsistent and ambiguous Software Requirement Specification (SRS) document results in an erroneous/failed software project. Hence, it is a serious challenge to handle and process complex and ambiguous requirements. Most of the literature work focuses on detection and resolution of ambiguity in software requirements. Also, there is no standardized way to write unambiguous and consistent requirements. The goal of this research was to generate an ambiguity-less SRS document. This paper presents a new approach to write ambiguity-less requirements. Furthermore, we design a framework for Natural Language (NL) to Controlled Natural Language (CNL) (such as Semantic Business Vocabulary and Rules (SBVR)) transition and develop a prototype. The prototype also generates Resource Description Framework (RDF) representation. The SBVR has a shared meaning concept that minimizes ambiguity, and RDF representation is supported by query language such as SPARQL Protocol and RDF Query Language (SPARQL). The proposed approach can help software engineers to translate NL requirements into a format that is understandable by all stakeholders and also is machine processable. The results of our prototype are encouraging, exhibiting the efficient performance of our developed prototype in terms of usability and correctness.

Reducing Requirements Ambiguity via Gamification: Comparison with Traditional Techniques

“…A tool for detecting lexical and syntactic ambiguity in words and sentences [ 65 ] used parts of speech (POS) tagging on corpus of ambiguous words including “always,” “every,” “none,” “certainly,” “therefore,” “good,” “often,” “rejected,” “such as, “they,” and “those.” But the tool does not provide detailed description of those words or sentences that were ambiguous or created ambiguity. Similarly, a framework for NL transition to controlled NL (CNL) was designed in a way to lessen the complexity of ambiguity formed in a sentence [ 66 ]. The aim of this work was to avoid inducting ambiguity in SRS document at first place.…”

“…But the tool does not provide detailed description of those words or sentences that were ambiguous or created ambiguity. Similarly, a framework for NL transition to controlled NL (CNL) was designed in a way to lessen the complexity of ambiguity formed in a sentence [66]. e aim of this work was to avoid inducting ambiguity in SRS document at first place.…”