How do you assess a software modules's quality? It is a question I have been struggling with for some time as I try to perform a peer review of a large code base.
Over time, a software module evolves from its intended form to something less than beautiful as bugs are discovered (and fixed) and enhancements over the original requirements are implemented. This is particularly true for code which is developed on a multi-person project, where personnel can change and often a module gets changed by different engineers. Although I adhere to the rule, that the code structure should reflect the original author's style (and how many people change the comment at the top of the file to identify that they have been one of the author's? This assumes that this information isn't automatically added by the configuration management system.), it can become increasingly difficult to make changes.
So what is the best way to assess code quality throughout it's development?
No comments:
Post a Comment