Requirements specification understandability evaluated with cohesion, context, and coupling.

摘要

A major factor influencing the cost and success of software systems development is the early detection of errors in the software life cycle. When Requirements Specifications have architectures that facilitate understanding, errors can be found and corrected, or prevented. Understandability is a requirements specification quality about which there has been little research. The state-of-the-art method of assessing a specifications for understanding is through the subjective feedback of experts. The flaws associated with this method of assessing specifications make it desirable to turn to an objective, automated method of assessment. The methodology developed for this research incorporates an automated tool to assess a specification for the strength of three characteristics: cohesion, context, and coupling.;The goals of this research were to demonstrate a positive correlation between quantitative cohesion, context, and coupling attributes and the understandability of a requirements specification; and to demonstrate the ability to automate the assessment of the three attributes. To this end, expert assessments were correlated with an automated tool's assessment of a test requirements specification. A preponderance of agreement was demonstrated to substantiate the belief that attribute assessment by objective, automated tools is possible and is correlated with understandability.