This may be true for math. This is definitely not true for other engineering majors. One component (though perhaps not as substantial of a part of the overhead as one would think) is the difference in salaries for instructors and professors in engineering majors. Engineering professors typically make 30%-40% more than their equivalent counterparts in other departments.

I have absolutely no idea who you are, so I hope this comes across only as gentle advice. I strongly suggest that you do not single out black history out of the variety of cultural histories (or just history in general) as an exemplar of a worthless major. There is no technical approach that can demonstrate that black history is worthless and given the sensitivity around black culture (FYI, I am black) you just can't win making a statement like this one. Of course that is, unless, your goal is to divide opinion versus making an educated technical argument.

This sort of awareness is the problem with the tech community. The rest of your post has a very reasoned solution. In fact, I would say you sound like a very sharp guy, perhaps even a technical badass. However, the only other community where badassery clearly trumps politics is finance. Nobody likes finance. Without a nurturing connection to society and its current non-technical concerns (no matter how irrelevant they may seem), finance has been relegated to ruining the world by working on problems that people have decided have no societal value. The tech community's path to maximum impact includes at least some basic understanding of how people from different cultures perceive our culture and changing our attitudes about statements like this would be a step in the right direction.

This has already been done, see Microsoft Research. MSR has received some serious scorn from traditional academics, especially in the systems and programming languages communities because, frankly, they are wildly successful in both scale (# of publications) and impact (they have produced some popular research and industry tools). The reason why is because they have more people, more money, and more time (no need to teach). The one thing they are missing is sustained cheap labor: students (read interns) typically spend no more than 3 months on a project.

It also comes in a configuration with 8GB.

What? Thrun didn't even go to Stanford until 2003. Not to knock Thrun, but to suggest that the department hadn't already gone down that path is an insult to quite a few equally amazing professors in the department who had been doing statistical methods before this arrival.

You are right. Entities can't cooperate without shared communication. And communication between entities is only as secure as the communication contract between them. Hence, it is entirely plausible for a hacker to compromise one entity and use its communication contract to compromise another entity--IF the contract allows for it.

For instance, the contract could still certainly allow an entity to pass arbitrary pieces of code for another entity to run or allow an entity to change an arbitrary memory location in another, but this would either be a design decision (and a poor one at that) or due to a vulnerability in the implementation of the communication contract. Address space isolation removes from the communication contract the ability to twiddle any/all of another entity's bits modulo the explicit implementation of this capability or vulnerabilities in the communication contract. The latter is certainly a concern, but now our trusted base is reduced to the design and implementation of the communication contract as opposed to the sum of functionality in both entities.

Address-space isolation for reliability and security is at least as old as Mach (25 years?) and it only claims to be more secure and not a panacea.

oops. posted anonymously