Object-oriented model refinement technique in software reengineering

Abstract

There are many legacy systems that are pooly structured, but contain substantial corporation knowledge, consequently it is difficult to throw them away and rebuild the systems. Software reengineering offers an exciting opportunity in migrating such legacy systems to evolvable systems in a disciplined manner. Also, object-oriented rearchitecturing is being considered promising in such legacy systems reengineering, because of many advantages offered by object-oriented technique. In twofold-strategy software reengineering, one of the problems to be solved is to find an object-oriented model from the output of reverse engineering and the output of forward engineering. In general, the outputs of forward and reverse engineering can be inconsistent in their abstract levels, the amount of design information, naming convention, and structure. In this thesis, we present an object-oriented model refinement technique(ORT) to build an object model from these inconsistent information. We first organize the information from reverse engineering as specification information tree, and then compare the entities in specification information tree with the information from forward engineering with tree-structured data dictionary and class dependency matrix. We demonstrate the usability of ORT by an example.