holtz@clan.UUCP (Neal Holtz) (04/01/84)
Our net link is going down, so before the rest of the world is isolated, here are some random thoughts on the issue of personal computers in undergrad education... 1. Yes, IBM PC's are probably adequate to teach initial undergrad computing the way it has been done in the past -- that simply isn't good enough. I find it simply incredible that our engineering students, for example, are not taught how to use a good source-level debugger. The point is that real machines (not toys like IBM PC's, Apples, etc.) can provide an environment where students can develop and understand large programs (the programs themselves may not exceed 64K, but the environment sure will). Minimum environment: - multi window display. - debugger that displays source code, properly. - full screen editor - integrated and used for all interactions. Sure its expensive, this year. But if our school gets anything less, it will be 5 or 10 years before we will be able to have this environment. 2. I am more interested in making more use of workstations in our undergraduate engineering program. There the requirements are much more extensive. A few engineering applications programs are required, but mostly we require good graphics, tools like TeX, Macsyma, etc. Remember that this a chance to *IMPROVE* our undergrad education, not to simply continue with the old ways. The point here is that a flock of lesser machines obtained for the single purpose of teaching first year computing makes it much less likely that I will be able to get the environment I want for more advanced work. 3. MIT EE freshmen understand and modify 50 page Lisp programs a few weeks after the start of the term. They are learning incredibly more about computing than our students who are still struggling with 100 line Pascal programs after a year. As said in 1., above, the environment provided can dramatically increase the complexity of programs that can be understood by first year undergrads. 4. In the last two or three weeks, a couple of people here of ported 100 to 150,000 lines of code (in C, Pascal, and Fortran) from various machines to some Apollos. Almost all of that is usable in some way in our undergrad program. The Apollo environment is quite good, and made that task relatively easy (and fun). There is *ABSOLUTELY NO WAY* that any of that code would be portable to PC's. Summary: 64K PC's may be adequate to teach toy programming, but they are certainly inadequate for any serious computing for undergrads. Unless you are sure that introduction of these things for first year courses will not harm your chances of getting real machines, fight their introduction with every breath.