|
Département d'ingénierie informatique |
Project leader : K. Mens
Researcher : D. Ordonnez
Collaborations :
- R. Wuyts (DeComp, ULB, Belgium),
- T. Mens (Service de Génie logiciel, UMH, Belgium),
- S. Demeyer (LORE, UA, Belgium),
- S. Ducasse (Software Composition Group, University of Berne, Switzerland),
- P. Klint, M. Van Den Brand (CWI, Amsterdam, The Netherlands),
- R. Lämmel (Microsoft, Redmond, USA),
- Rhea System.
Description :
Related to our research on intensional views which provide a high-level description of the structure of a software system, we are interested in techniques that may help us to automatically extract architectural descriptions from the source code of a software system :
- The technique of formal concept analysis or clustering techniques may be used to mine the source code automatically for groups of source-code elements that have similar characteristics. Note that this approach is very similar to the approach we use for mining the source-code for potential aspects.
- Concept learning techniques, and more specifically learning by induction, may be used to automatically induce the 'intension' (i.e., a formal description of the commonalities of the elements) of an explicitly enumerated set of source-code entities (for example, a set that has been extracted from the source-code by means of formal concept analysis). See for example the paper on inductive logic programming where the technique of inductive logic programming is used to automatically induce intensional views from extensionally defined ones. The same technique may also be useful to automatically induce relations among intensional views.
- Description logics may offer a suitable underlying formalism for intensional views. Since they have a restricted syntax and specific reasoning capabilities, they may make it easier to extract intensional views and intensional relations that are described in such a logic.
In the context of our research project APPAREIL we will investigate a reengineering approach that is as independent as possible of the programming and design language being considered and apply it to the MOIS environment of our industrial partner Rhea Systems. We intend to combine our expertise on logic meta programming with generic language technology that has been developped at CWI.
In the context of our FRFC-funded research center on software restructuring (with ULB and UMH), we will study formalisms for automated software restructuring, refactoring techniques, transformation approaches and the model of intensional views.