Currently, software systems must be appropriately developed to support an amount of variability for accommodating different requirements. To support such development, a diversity of tools has already been designed for variability management (i.e., identification, modeling, evaluation, and realization). However, due to this diversity, there is a lack of consensus on what in fact software variability tools are and even what functionalities they should provide. Besides that, the building of new tools is still an effort- and time-consuming task. To support their building, we present VMTools-RA, a reference architecture that encompasses knowledge and practice for developing and evolving variability tools. Designed in a systematic way, VMTools-RA was evaluated throughout: a controlled experiment with software developer practitioners; and an instantiation of the VMTools-RA architecture to implement a software variability tool, named SMartyModeling. As a result, VMTools-RA is evidenced to be feasible and it can be considered an important contribution to the software variability and developers of variability-intensive software systems community, which require specific tools developed in a faster manner with less risk, what a reference architecture could provide.

software variability, reference architecture, SMartyModeling, VMTools-RA, tool, variability management