Experiments in identifying reusable abstract data types in program code

The issue of program comprehension is addressed from the software reuse perspective. In particular the identification of abstract data types in existing program code is explored. A candidature criterion is presented and a prototype implementing it is described. The criterion is applied in an experiment that analyses five very different programs and the resulting output is discussed. The work described forms part of the RE/sup 2/ project that addresses the wider issues of software reuse through the exploration of reverse engineering and re-engineering techniques to identify and extract reusable assets from existing systems.<>