A Heuristic Approach to Detecting Cross-Language Links

Zusammenfassung:

Most software systems today are so-called Multi-Language Software Systems or Applications (MLSS/MLSA), that is, they are written using more than one programming language. The artifacts in different languages are linked, at the latest at runtime, forming cross-language links.

A common scenario for developing MLSAs is using one general-purpose language – such as Java – and a multitude of DSLs for the individual tasks at hand.

In many cases, such DSLs reference Java artifacts like classes or methods by name. Thus, cross-language links can be detected by searching for the Java artifact names in DSL code and applying heuristics to determine which artifact, if any, is linked.

This thesis investigates such an approach and presents various implementation strategies. Precision and recall are tested on a case study.

Download:
N/A