lord+@andrew.cmu.edu (Tom Lord) (04/17/88)
Peter Su (hugo@eleazar.Dartmouth.EDU) observes that CMU requires a programming course of all students, and complains that it is in Pascal rather than LISP, Scheme, or some other interpreted language with a simple syntax. First, for the record, the requirement to which Peter refers can now be satisfied by a `gentle' LISP course. The course is still almost universally condemned as useless by mathematics, engineering, humanities, and fine arts majors alike. Also required of (most or all) students, is a course called "Computing Skills Workshop". This course is an introduction to word processing, email and electronic bboards, and various other computing facilities available to CMU students (such as the excellent on-line encyclopedia service). This course is generally well received. It's most tangible biproduct is that our public computer clusters are filled with a higher number of non-technical students than clusters I have seen at other universities. I believe (and I think many students do, too) that the workshop really enhances the education of most undergraduates here. Finally, I would like to raise a new, though related issue. Here at CMU, where most if not all students are required to take at least one programming course, it is surprising how LITTLE programming a typical computer science major gets away with. As far as I can tell, it is possible to graduate with a BS in Applied Mathematics, Computer Science having worked on only one program substantially longer than a page or two. It seems to be typical to graduate having only worked on two such programs. (The degree is an applied math degree because the CS department has not been able to devise its own undergraduate major. They've talked about it an aweful lot though :-) ) Now, regardless of your opinions about "What is CS?", I hope that you'll find this situation apalling. (1) This situation obviously can't produce high quality production programmers. (2) Less obviously, it can't produce very good researchers - since a common failling of research projects is to abandon the practices of quality production coding. Finally (3), I doubt that this program does much to help produce theoriticians, since as many have pointed out (more or less), the best theory is that which takes the real world into account, rather than scoffs at it. I have three questions: First, am I just wrong? Assuming that the situation is as I described, will quality computer scientists emerge from such a program? Better than if the program required more actual programming experience? Two, are things the same at MIT, Stanford, and other `high-power' computer science schools. Three, (begging a "no" answer to the first question) how do you repair the problems here? Thomas Lord lord+@andrew.cmu.edu