Building empirical software engineering bodies of knowledge with systematic knowledge engineering
Stefan Biffl, Marcos Kalinowski, Fajar J. Ekaputra, Estefanía Serral, Dietmar Winkler · Software Engineering and Knowledge Engineering · 2014
[Context] Empirical software engineering (EMSE) researchers conduct systematic literature reviews (SLRs) to build bodies of knowledge (BoKs). Unfortunately, valuable knowledge collected in the SLR process is publicly available only to a limited extent, which considerably slows down building BoKs incrementally. [Objective] In this paper, we introduce the Systematic Knowledge Engineering (SKE) process to support building up BoKs from empirical studies efficiently. [Method] SKE is based on the SLR process and on Knowledge Engineering (KE) practices to provide a Knowledge Base (KB) with semantic technologies that enable reusing intermediate data extraction results and querying of empirical evidence. We evaluated SKE by building a software inspection EMSE BoK KB from knowledge acquired by controlled experiments. We elicited relevant queries from EMSE researchers and systematically integrated information from 30 representative research papers into the KB. [Results] The resulting KB was effective in answering the queries, enabling knowledge reuse for analyses beyond the results from the SLR process. [Conclusion] SKE showed promising results in the software inspection context and should be evaluated in other contexts for building EMSE BoKs faster. Keywords-Empirical software engineering, systematic knowledge engineering, systematic review, software inspection.