Comprehending source code of large software system for reuse
Aniket Kulkarni · 2016
Comprehending source code of large software system for code reuse is an important problem. It is due to very high cost of software maintenance. In this paper, the author shares the experience on comprehending code-base of PRISM - a model-based program analysis tool developed by Tata Consultancy Services - containing in total more than 500 KLOC. During this activity, the author not only detected two bugs in existing PRISM software system while comprehending more than half of the total code (approximately 300 KLOC), but also added feature of Application Programming Interface (API) for Intermediate Representation (IR) transformation utility for `C' language consisting of 1 KLOC. The author applied combination of key techniques such as lexical and dynamic analyses, models such as Brook, Soloway to understand code-base and, mixed top-down and bottom-up approaches for comprehension. The author observed that speed of comprehension and reuse later increased multi-fold because half of code-base (approximately 250 KLOC) is automatically generated from specification and thus by focusing on understanding less than 25 KLOC.