This post on why academics should work to integrate their research ideas with open source software in order to actually push forward on the OSS side (and presumably vice versa) is really good. I think such things could be applied in software testing, too, where there's a gulf between how academics think and what people actually use. I was impressed with Dr. Tao Xie's work on PEX and Fitnex with respect to testing, and I'm talking with him about applying it to Python.

--titus