Greg Wilson has an interesting post regarding both the lack of evidence and the lack of regard for evidence that is prevalent in software engineering.
He mentions a commenter that says “As a working programmer I favour my own personal experiences and experimentation over academic studies…”. This widespread (though not universal) opinion makes software development one of the few modern professional fields where practitioners still largely shun scientific evidence.
To their credit, it’s also still hard to find relevant, trustworthy, replicated empirical research in our field. Greg points towards Glass’s excellent Facts and Fallacies of Software Engineering and to the Empirical Software Engineering Journal (semi-restricted access). Steve McConnell plays an excellent role at this as well. But we need many more bridges connecting industry and academia, for the benefit of both communities. Scientific software engineering blogs, anyone?