Abstract
The identification of objects in procedural programs has long been recognised as a key to renewing legacy systems. As a consequence, several authors have proposed methods and tools that achieve, in general, some level of success, but do not always precisely identify a coherent set of objects. We show that using an eclectic approach, where a domain expert software engineer is encouraged to tailor and combine existing approaches, may overcome the limitation of the single approaches and helps to better address the particular goals of the project at hand and the unique aspects of the subject system. The eclectic approach is illustrated by reporting experiences from a case study of identifying coarse-grained, persistent objects to be used in the migration of a COBOL system to a distributed component system.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
