A Metric-Based Approach to Detect Abstract Data Types and State Encapsulations

  • Authors:

  • Jean-François Girard;Rainer Koschke;Georg Schied

  • Affiliations:

  • Fraunhofer Institute for Experimental Software Engineering, Sauerwiesen 6, D-67661 Kaiserslautern, Germany. girard@iese.fhg.de;University of Stuttgart, Breitwiesenstr. 20-22, D-70565 Stuttgart, Germany. koschke@informatik.uni-stuttgart.de;University of Stuttgart, Breitwiesenstr. 20-22, D-70565 Stuttgart, Germany. schied@informatik.uni-stuttgart.de

  • Venue:
  • Automated Software Engineering
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

This article presents an approach to identify abstract datatypes (ADT) and abstract state encapsulations (ASE,also called abstract objects) in source code. This approach, namedsimilarity clustering, groups together functions, types, andvariables into ADT and ASE candidates according to the proportion offeatures they share. The set of features considered includes thecontext of these elements, the relationships to their environment,and informal information. A prototype tool has been implemented tosupport this approach. It has been applied to three C systems (eachbetween 30–38 Kloc). The ADTs and ASEs identified by the approachare compared to those identified by software engineers who did notknow the proposed approach or other automatic approaches. Within thiscase study, this approach has been shown to have a higher detectionquality and to identify, in most of the cases, more ADTs and ASEsthan the other techniques. In all other cases its detection qualityis second best. N.B. This article reports on work in progress on thisapproach which has evolved since it was presented in the originalASE97 conference paper.