Quality Assessment Method for Software Requirements Specifications Based on Document Characteristics and Its Structure

Abstract

The Software Requirements Specifications document (SRS) is used to collect the user requirements, which is used as an input for development process, and as a baseline for verifying the correctness of the software product occurring at each step throughout the software development process. It has been found that many organizations cannot deliver software products that satisfy the actual requirements of the customers, due to defects that frequently occur in the SRS, especially the use of ambiguous natural language in the requirements specifications and the inappropriate document structure, which negatively affects the software quality. Therefore, this research presents a method for objectively assessing the quality of the SRS for use of natural language in requirements specification, document structure, and overall document quality, by directly considering three characteristics of the document: unambiguous, verifiable and modifiable, to indicate the quality of the document and defects that appear during the software requirements engineering step. The process assessment model is applied as a framework for assessing the quality of the SRS, and the measurement process model and measurement information model are used as approaches for proposing a method for SRS quality assessment and defining the metric, respectively, using Pearson's correlation coefficient as the criterion for verifying the validity of the results obtained from assessing the SRS using the proposed method, indicating that the results obtained from the quality assessment reflect the quality of the SRS, as well as the apparent defects.