DagSemProc.06302.7.pdf
- Filesize: 458 kB
- 10 pages
As software evolves, new functionality sometimes no longer aligns with the original design, ending up scattered across a program. Aspect mining identifies such cross-cutting concerns in order to then help migrating a system to a better design, maybe even to an aspect-oriented design. We address this task by applying formal concept analysis to a program's history: method calls added across many locations are likely to be cross-cutting. By taking this historical perspective, we introduce a new dimension to aspect mining. As we only analyse changes from one version to the next, the technique is independent of a system's total size and scales up to industrial-sized projects such as Eclipse.
Feedback for Dagstuhl Publishing