Experimental Comparison of the Comprehensibility of a Z Specification and its Implementation

Acknowledgements The authors wish to acknowledge the support of UK EPSRC, which has funded the Empirical assessment of formal Methods (EMPAF) project, (GR/L87347) and a PhD studentship for this work. We would also like to thank the students who participated in the experiment. Abstract Comprehensibility is often raised as a problem with formal notations and yet practitioners using formal methods dispute this. In a survey, one interviewee said "formal specifications are no more difficult to understand than code". Measurement of comprehension is necessarily comparative and perhaps the most useful comparison for a specification is with its implementation. Practitioners have a reasonably good intuitive feel for the comprehension of code and a quantified comparison will therefore transfer this feeling to formal specifications. We performed an experiment to compare the comprehension of a Z specification with that of its implementation in Java code. The results indicate that there is little difference in comprehensibility between the two.