Michael Mitzenmacher, a CS professor at Harvard, asks this question. I found many of the responses interesting, several which echo comments I've made on the subject.
I've actually found his blog, My Biased Coin, fascinating reading. He touches on quite a few issues I've discussed here, such as the importance of theory in CS education, and how exams should be designed. He even gives some comments on Ben Edelman's work on click fraud and related topics. The only topic I haven't seen covered is immigration/visas; I'm interested in what he has to say. As another point of interest, he worked with Andrei Broder, who was Chief Scientist of AltaVista many years ago, and is now part of Yahoo!'s research group. I'll definitely be checking his blog and commenting on what I read both there and here.
I've actually found his blog, My Biased Coin, fascinating reading. He touches on quite a few issues I've discussed here, such as the importance of theory in CS education, and how exams should be designed. He even gives some comments on Ben Edelman's work on click fraud and related topics. The only topic I haven't seen covered is immigration/visas; I'm interested in what he has to say. As another point of interest, he worked with Andrei Broder, who was Chief Scientist of AltaVista many years ago, and is now part of Yahoo!'s research group. I'll definitely be checking his blog and commenting on what I read both there and here.
- Current Mood:
impressed
Profile
gregbo- gregbo
- gregbo's homepage
Comments
Personally, I think that MIT in general taught me how to be resourceful and deal with stress, and being course 9, with its extremely interdisciplinary nature, taught me how to learn the basics of any subject that I needed quickly (and also, it taught me Matlab). 6.170, in addition to forcing me to learn Java reasonably well, taught me enough about software engineering practices to make a good impression on interviewers. Something like 6.004 was less directly useful, but it was still a great class. I sometimes use the probability & statistics that I learned in 9.07, at work.
In the post-bac program at Tufts, I think I've had a good balance of classes. Computer Graphics sure improved my C++ skills, which were directly applicable to work, because we did do significant programming assignments and a substantial individual final project. Algorithms, I use occasionally. Currently, I'm in Programming Languages, which has both programming and non-programming assignments, and we've done programming assignments in Scheme and ML (and will apparently have one in Prolog). I know of at least one company that uses Scheme, but I'm unconvinced of the worth of knowing ML either for instructional or industry purposes.
Of course, eventually I want to get a PhD and do research, so my career goals are not the same as someone who wants to be a software or network engineer. Which is an issue that one of the commenters addresses - people have different career goals, and why should the future programmers and engineers be served at the expense of the future scientists (course 9's sort-of-parallel was premeds vs. non-premeds, I guess)?