An Experimental Evaluation of Documentation Methods and Reusability

This paper presents an experimental evaluation carried out in an academic environment. The goal of the experiment was to compare how different methods of documenting semantic information affect software reuse. More specifically, the goal was to measure if there were any differences between the methods with regard to the time needed to implement changes to existing software. Four methods of documentation were used; executable contracts, non-executable contracts, Javadoc-style documentation and sequence diagrams. The results indicate that executable contracts demanded more time than the other three methods and that sequence diagrams and Javadoc demanded the least time.