An experiment in identifying persistent objects in large systems

We present an experiment in identifying coarse-grained persistent objects in a legacy system of an Italian public organisation. Object methods are searched for at the program level driven by the minimisation of the coupling between objects. This strategy is useful in incremental migration projects requiring the identification of largely independent subsystems needing low re-engineering and decoupling costs to be first encapsulated in different wrappers and then selectively replaced. The aim of the experiment was to evaluate the feasibility of this approach when applied to large software systems. The work presented in this paper is part of the project PROGRESS, a research project on process and software re-engineering in Italian public organisations carried out by Italian universities and research centres.