> But suppose not. Even if this isn't industrial espionage, there's
> still the possibility that Kaashoek is just using his class at the
> university as unpaid R&D for his own web-caching startup. Would
> anyone else have a problem with this?
Sure, but in reality it's a logical extension of past and current Institute
behavior of ripping off the IP of students. It actually sets up some very
a) Assuming we're talking about a class assignment between prof and student,
the student still owns the IP (at least under current MIT policy). It's not
clear the student has a cause of action against Kaashoek-the-prof for simply
turning around and giving the idea to Kaashoek-the-startup. Were
Kaashoek-the-prof to hand over a physical copy of the assignment, then you'd
likely have both a copyright violation and also a Buckley Amendment (20 USC
1232g) violation, too. Of course, the startup couldn't file for a patent on
the idea without bringing in the student, so the student would have an
opportunity to bargain at that point.
b) Were the student on MIT support (e.g. grad student RA), it is current
Institute policy to claim all the IP of the student as work-for-hire. Then
you'd have an IP fight between MIT and the start-up unless some prior
contract existed between those two parties but the student would be SOL.
When you think about it, this is really no different than the commonplace
(at least within LCS) practice of forcing students to scribe notes in grad
classes, them turning around and using those notes as the basis of a
textbook. It's still an institutionalized theft of IP, but the student is
essentially powerless to fight it because (a) he needs the diploma, and the
Institute has complete control over that, and (b) the Institute has infinite
resources to fight any attempted legal action.
As a co-worker of mine (who also spent way too long at LCS/AI) said to me
today when I told him about the WSJ catfight story, "Couldn't happen to a
more deserving group of people."