Evidence-based software engineering revisited: evaluation of a practice-driven application.

Abstract

Systematic reviews (SRs) are the main method for supporting evidence-based software engineering (EBSE). However, although SRs have been widely adopted by SE researchers, so far, there have been no studies assessing the value of the broader EBSE framework. Objective: To evaluate an EBSE application in an industry environment. Method: Using the participant observation method, we conducted an EBSE-based project to address an industry problem. This includes collaborating with practitioners to diagnose the problem, collect evidence through a rapid review (RR), and transfer the results to the company. Results: The practitioners utilized some recommendations that addressed their problems. The biggest barriers we encountered were the difficulty in finding relevant evidence, the complexity of applying evidence, and the lack of guidelines or examples of EBSE use. Factors that supported our project were: close collaboration with the company and its commitment to process improvement, appropriate dissemination of the results, using an RR, and participation of external researchers. Conclusions: Currently, the use of EBSE is challenging and requires both professional and research skills. While it is a valuable instrument for researchers to enhance collaboration with industry, it may not be as suitable for general use by practitioners, as initially hoped.