Validation of software measures for the functional programming language Erlang
Daniel Warmuth · edoc Publication server (Humboldt University of Berlin) · 2019
Through the use of software measures, relevant properties of the processes and products of software development are to be quantified and objectively assessed. The validation of measures aims at demonstrating that measures actually capture those properties that they are targeted at. In view of the increasing importance of functional programming and the associated expectations, like improved maintainability and other aspects of higher software quality, the present study examines the empirical validation of software measures in the domain of functional programming. The studies published to date on the validation of software measures for functional programming languages are based on data from small or medium-sized experimental software projects with no more than a few thousand lines of code. For this reason, the present study considers a larger software project which contains tens of thousands of lines of code and is in constant professional use: the widespread communication server ejabberd, which is implemented in the functional programming language Erlang. Various software measures are obtained and the relationships to external measures of quality are studied; these measures are derived from a data base of known issues and from code modifications. On this basis, statements are made as to the extent to which specific software measures can be used as indicators for specific properties of a program. Further, comparative values which may be useful for evaluating the measurement values for other systems are determined.