Feature models are a key element in software product lines, representing the supported features and their interrelationships within a family of software products. Recommendation systems for software engineering (RSSEs) are potentially useful in supporting the extraction, maintenance, and categorization of feature models. This paper focuses on the design and implementation of an RSSE to automatically recommend features for software product lines, the types of these features, and how they could be related to each other. Such a recommender should save time and tedium over doing the work manually. We present FFRE, a prototype recommendation tool for the extraction of features and their relationships from software requirements specification (SRS) documents. FFRE is based on natural language processing (NLP) techniques and heuristics. FFRE is evaluated qualitatively from four SRS documents and compared against other tools and approaches.
The importance of having mature software development methodologies and tools for the increasingly popular pervasive systems cannot be understated. Focusing on system architectures, we previously conducted a thorough review of over 50 state of the art architectures related to pervasive systems. From the review, we elicited a set of major features that should be supported in pervasive systems, along with best practice architectures for designing such features. We then detailed a methodology, through which designers of new pervasive systems can select a set of desired features and generate a baseline architecture for their system. In this article, we evaluate our methodology with an empirical study that compares generated architectures with ones designed by subject matter experts with sufficient experience in the domain. We used different evaluation suites and measurement techniques in our comparisons. Results show that our automatically generated architectures are very comparable with, and in many cases of higher quality than, the architectures designed by subject matter experts.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.