Clone detection in telecommunications software systems : a neural net approach

We report on the development of a tool which uses neural nets to help in the detection of 'clone' software in large telecommunication software systems. Cloning is a primitive form of software re-use, whereby existing blocks of source code are copied and adapted, for use elsewhere in the system or to implement new functionality. A number of representation schemes for presenting the source code to Self-Organising Maps (SOMs) have been tried. The more favourable of these have been applied in a hybrid, modular tool architecture. The current result is a demonstration system in trial.