Other people have thought about, and at least started to solve the problem of academic source code being extremely proof-of-concept rather that production or resume ready pieces of software engineering art:

http://matt.might.net/articles/crapl/