What Makes A Research For Software Engineering - 1241 Words

The author, Shaw, tries to identify what are some of the better research strategies for research in software engineering in her paper titled, â€Å"What makes good research in software engineering?† By formalizing what is already considered as high quality research, software engineers can better design research processes, aide them in writing quality research papers and also help them read and better understand software engineering research papers. Shaw puts forth the hypothesis that research in software engineering should be characterized by research processes in other science and engineering disciplines which include research questions, results and validation. Highlighting related work, she mentions previous research efforts to characterize†¦show more content†¦Expanding on the model types which serve as results for the research questions, Shaw explains that the analytical model is a structural model providing support for formal analysis, the descriptive model exposes hurdles in decision making in the design phase and the empirical model which is uncommon provides sound statistical data to back it up. The research results proposed are validated using results based on controlled experiment or empirical model. Results validated through experience are generally not accepted. Having identified the three main characteristics of software engineering research, Shaw proposes that research strategies should be created as a combination of the three research characteristics. Shaw validates her proposal by pointing out some of the most common research strategies, as seen in widely accepted publications, follow this template of combination of research questions, results and validation. Some of the most common strategies include proposing new methods for software development which are validated through use cases and analyzing software development through analytical model, which are validated using formal analysis or experience. The author also argues that all major breakthroughs have been built on multiple papers with not enough substantial evidence. Thus, she mentions that software engineering research papers are incrementally progressive, suggesting that