I've been mulling this over the last few days.
I want to teach a class, but a very specific class. I want to teach a class that would train students to be engineers that I would hire and expect, right out of the gate, that they knew how to do the work required, or at least how to work with a team. A lot of my work in graduate school was isolated from the nine-tenths of work necessary to make a product commercial. For instance, I learned how to make a gaussian blur filter, and even learned how to make equivalent versions as a PDE, using the FFT, and as a straight-up filter, but I never learned how to wrap that knowledge into something that could be presented to the end user.
The thing is, in order to teach at a university in the US (and maybe in other parts of the world, I haven't checked), you have to publish papers on the field. Publication has nothing to do with industry success; in fact, I'd argue that they are inversely indicative of commercial success. Academic code only has to work on a small subset of data, and may not have been thoroughly tested and almost certainly hasn't been looked at in detail by anyone other than the author. That makes for very non-robust code, not code you can just roll out into the world. I want to teach students how to make production-level code, and that means looking at things like source control, software engineering, bug tracking, and so forth. Now, how to get a university to bite, when I don't want to publish? That's the question...
Thursday, March 18, 2010
Graduated!
I don't really update this much, but maybe I should?
Anyway, I'm all graduated now! Up next: What next?
Anyway, I'm all graduated now! Up next: What next?
Subscribe to:
Posts (Atom)
