An automated software transplantation approach for reengineering of systems into product lines

Abstract

Although Software Product Lines (SPL) offer the potential for order-of-magnitude improvements in software engineering performance, the up-front cost, level of effort, assumed risk, and latency required to make the transition to SPL are prohibitive adoption barriers for many organizations that could otherwise benefit from reusing of their existing systems. The SPL adoption from an extractive model based on a reengineering process of exist- ing systems into SPL is an active research topic with real benefits in practice. It allows software development companies to preserve their investment and aggregate knowledge obtained during the development of their portfolio of systems individually developed. Despite these benefits, adopting an extractive product line approach still requires a con- siderable upfront investment and is more complex to evolve than single systems. Because of these drawbacks, software companies refrain or delay the adoption of SPL, resorting to an ad-hoc practice of clone-and-own. To speed conversion to and maintenance of SPL, we present Foundry, a Software Transplantation (ST) approach that guides the process of transplanting and merging features in a product line from existing systems. It is the first approach for SPL that automates all stages of product line construction using the ST technique. We realized Foundry in prodscalpel, a software transplantation tool for SPL that automates the process of identifying, adapting and transferring features from existing systems to a common product base. Its code transfer mechanism between different systems allows it to be used not only for the generation of product lines but also as an alternative to the clone-and-own technique for system specialization. We compared our proposal with the existing reengineering solutions to demonstrate evidence that the ST is an alternative with potential for application in the field of reengineering of existing systems to SPL. In the search for more concrete evidence, we evaluated prodscalpel on two case studies where two products were generated by transplanting of features from three real-world systems. Moreover, we conducted an experiment comparing Foundry’s feature migration with manual effort. We show that Foundry automatically migrated features across codebases 4.8 times faster, on average, than the average time a group of SPL experts took to accomplish the task. Although preliminary, our evaluation provides evidence to support the claim that ST for Software Product Line Engineering (SPLE) is a feasible and, indeed, promising new research direction.