This is fantastic. The author is absolutely correct: academic software is crap. I've used software written by someone who wrote a well known tutorial on Haskell

This is fantastic. The author is absolutely correct: academic software is crap. I've used software written by someone who wrote a well known tutorial on Haskell, and his code was a pile of junk. This is the right way to do it in academia -- you are only rewarded for publishing so good academics write the minimal system necessary for publication and then dump the code. As a side note, Matt Might's backups suck if he could only find that one program ;) As another side note, the PLT group behind Racket are exceptional for publishing lots and writing good code.