Identification of green, yellow and red legacy components

Abstract

Software systems are often getting older than expected, and it is a challenge to try to make sure that they grow old gracefully. This implies that methods are needed to ensure that system components are possible to maintain. In this paper, the need to investigate, classify and study software components is emphasized. A classification method is proposed. It is based on classifying the software components into green, yellow and red components. The classification scheme is complemented with a discussion of suitable models to identify problematic components. The scheme and the models are illustrated in a minor case study to highlight the opportunities. The long term objective of the work is to define methods, models and metrics which are suitable to use in order to identify software components which have to be taken care of through either tailored processes (e.g. additional focus on verification and validation) or reengineering. The case study indicates that the long term objective is realistic and worthwhile.