City Research Online

Abstract

In this work our aim has been to investigate the application of software metrics to the area of formal specification. We first look at the nature of Formal Methods and those most commonly in use. This suggests the scope for the type of specifications that should come under investigation. Metrics and the nature of measurement become an important consideration as we discuss how to assess Formal Methods themselves and their impact on software development. We look at the work already done in software metrics to see if parallels can be drawn with the proposed work on formal specification. To establish current knowledge about the use and effectiveness of Formal Methods we investigate to see how their benefits had been assessed by two major surveys. Further information on the impact of the methods is obtained by analysing the results from a series of case studies. We build on work done in a previous study to investigate and compare attributes of three formal specification notations. This gives valuable insights into possible attributes, the way measurements might be made and possible metrics to establish. Drawing together the knowledge and experiences gained from the background research we design three experiments to look at metrics for comprehensibility in formal specifications. The experiments together with their results and statistical analysis are described. The main findings show that the variable names, comment levels and structure of Z specifications have an effect on their comprehensibility.

Publication Type:	Thesis (Doctoral)
Subjects:	Q Science > QA Mathematics > QA76 Computer software
Departments:	School of Science & Technology > Computer Science > Software Reliability School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads