Semantic Werks

I’ve wondered before why few commercial developers seem to care about academic results and tools. For instance, most of the tools I’ve worked on have … well, flown beneath the commercial radar is an apt metaphor.

Partly this is because not being professionals, our software lacks polish and the ‘last-mile’ effort needed for widespread adoption (suggesting open-source is probably the best venue for it).

However, Jorge Aranda had a great insight today about this issue. He mentioned a paper by Roel Wieringa, Neil Maiden, Nancy Mead, and Collette Roland (cost-wall protected, unfortunately) in the RE journal. Jorge’s take on the article was that since requirements engineering researchers often do not do empirically grounded research, there is little evidence for commercial practitioners to use in making adoption choices. To a lesser extent the same is true of software engineering in general, I think. That is, although researchers (and practitioners too) are quick to come up with new techniques, it is often years before any validation of those techniques is performed.

I contrasted this with materials science, or civil engineering. There, if a researcher has an idea for a new cement formulation, he or she does an empirical study (clearly easier in that domain). Armed with the evidence, commerical adopters move quickly to take advantage.

This isn’t the whole adoption story, of course. In particular the book “Diffusion of Innovations” makes clear that technical prowess, proven or not, is not a sufficient criteria for adopting something.

Still, I think adding empirical validation in this area of research is a positive — and very necessary — contribution to the field.