This paper addresses the problem of reengineering object-oriented systems that have incurred increased maintenance cost due to long development time-span and project lifecycle. When an Incremental Approach is used to develop an object-oriented system, there is a risk that the class design and the overall object model will deteriorate in quality with each increment. A recent research work suggested a process activity (Class Deterioration Detection and Resurrection - CDDR process activity) and a technique for the detection and resurrection of deteriorated classes [5]. That work focussed on one particular aspect of object-oriented software maintenance - Class Quality Deterioration due to lack of cohesion induced by high coupling. This paper addresses the problem of deteriorating object-oriented design due to code and class growth (increase in the number of classes) within a system. A Code/Class Growth Control process activity (CGC) is suggested to avoid and eliminate Repetitions Code and Classes within the evolving system. The CDDR and CGC process activities are used to build an evolving Maintenance process model for object-oriented systems. The presented maintenance process model is an effective way to periodically assess and resurrect the quality of an object-oriented design during incremental development.

When an Incremental Approach is used to develop an object-oriented system, there is a risk that the class design will deteriorate in quality with each increment. This paper presents a technique for detecting classes that may be prone to deteriorate, or if deterioration has occurred assists with reengineering those classes. Experience with applying this technique to an industrial software development project is also discussed.

The practice of software development continues to shift towards the use of object oriented approaches. The motivation for this trend is the benefits attributed to object oriented software, including improved maintainability. As organizations develop new object oriented software, they face the problem of maintaining their older software. How can existing non objected oriented software benefit from this new software engineering technology? The paper presents a nine step process for deriving an object model from existing unstructured FORTRAN source code. Both top down and bottom up approaches are used to derive objects, classes, class attributes and methods, and relationships among classes. This process can be used within a reengineering project to convert legacy FORTRAN code into a new object oriented implementation written in a language such as C++. Experience with using this process is also described.