Professor John Regehr's Closing the Loop: The Importance of
External Engagement in Computer Science Research is an excellent article on
the general spots where academic computer science can become
disconnected with the real world and the engineering problems that
are found there. Since I work in academia (and have read Greg
Wilson for some time), this is an issue
relatively near to my heart and I quite liked how he presents things
in the article. It's a new framing of the issues, one that puts
things in a clear way.
He's also written a followup post, Paths to External Engagement
in Computer Science Research.
This one is probably mostly of interest to people inside the sausage
factory who want to interact with the outside, as opposed to people
on the outside wondering why on earth academic computer science isn't
more useful to them.