Towards a Process for Developing Maintenance Tools in Academia

Building of tools---from simple prototypes to industrial-strength  applications---is a pervasive activity in academic research. When  proposing a new technique for software maintenance, effective tool  support is typically required to demonstrate the feasibility and  effectiveness of the approach. However, even though tool building is  both pervasive and requiring significant time and effort, it is  still pursued in an ad hoc manner. In fact, little research has  addressed the question how to make tool building in academia more  disciplined, predictable and efficient.  In this paper, we address these issues by proposing a dedicated  development process for tool building that takes the unique  characteristics of an academic research environment into account. We  first identify process requirements based on a review of the  literature and our extensive tool building experience in the domain  of maintenance tools. We then outline a process framework based on  work products that accommodates the requirements while providing  needed flexibility for tailoring the process to account for specific  tool building approaches and project constraints. The work products  are concrete milestones of the process, tracking progress,  rationalizing (design) decisions, and documenting the current state  of the tool building project. Thus, the work products provide  important input for strategic project decisions and rapid initiation  of new team members.  Leveraging a dedicated tool building process promises tools that are  designed, build, and maintained in a more disciplined, predictable  and efficient manner.