Smell Detection in Context.

Daniel Speicher, Sebastian Jancke · 2010

Good design is essential for long-term evolvability of software systems. To reduce the required additional short-term effort appropriate tools are highly desirable. The most promising approach [3] combines basic metrics to build complex heuristics to detect refactoring opportunities (design smells). We still saw-backed by a questionnaire survey in [5]- the need to systematically improve this approach. The problems we want to address are: 1. Insufficient handling of false positives 2. Information overload (a) Irrelevance in the context of the current task (b) Irrelevance because of historic stability But how can a tool- developed based on a lot of expertise- be wrong anyway? We see the main reason

Read the paper · More papers on PaperTik