liberte@uiucdcsb.CS.UIUC.EDU (09/23/86)
How much of a numerical methods course is a CS person likely to use? Granted there are a few very important principals that everyone should know about. But maybe they should be included in a basic programming course. Unless you are going to apply your CS skills to a field of engineering or some area that actually uses numerical methods, or unless you want to do research in numerical analysis, I don't think it helps much to take a whole course in the subject. My original argument was not that Numerical Methods isn't important - every field IS important - but that it should be put on a more equal basis with other application areas. A student should be required, for example, to take two or three courses out of the several application (or peripheral) areas. Other peripheral areas include data bases, graphics, AI, and perhaps operating systems and digital design. I think every CS person should be familiar with all of these areas (and a few others), but to require a whole course in each of them is over doing it. So perhaps there should be a one year required survey course similar in structure to the survey courses in other departments. It would touch on each area in CS for two or three weeks just giving the basic principals that everyone should know about. After that, a student would be in a better position to choose which areas to pursue further. Other required courses should include programming, data structures (and algorithm structures really), theory of computation, and probably software engineering since just about every CS person does programming. I am enjoying this discussion despite the sometimes inflammatory remarks. I can see why people may become inflamed considering the years of time invested in education. But let's hold onto our thinking caps. Dan LaLiberte liberte@b.cs.uiuc.edu liberte@uiuc.csnet ihnp4!uiucdcs!liberte