Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!tektronix!uw-beaver!ubc-vision!alberta!cdshaw From: cdshaw@alberta.UUCP (Chris Shaw) Newsgroups: net.cse Subject: Re: CS vs. CE Message-ID: <65@alberta.UUCP> Date: Mon, 8-Sep-86 19:33:47 EDT Article-I.D.: alberta.65 Posted: Mon Sep 8 19:33:47 1986 Date-Received: Thu, 11-Sep-86 19:10:34 EDT References: <13500008@uiucdcsb> Reply-To: cdshaw@pembina.UUCP (Chris Shaw) Organization: U. of Alberta, Edmonton, AB Lines: 80 In article liberte@uiucdcsb.CS.UIUC.EDU writes: >Regarding an earlier note on electrical engineering: > >The question of how much electrical engineering is relevant to >computer science study depends on how you define the fields. >Since there is a field of computer engineering as distinct from >computer science, we should take advantage of the fairly distinct >boundary between these fields. This is silly. There is barely enough material to cover a hardball 4-year honours undergraduate degree as it is, without chopping the field into two totally different disciplines. Imagine if you were to do the same with chemistry. Would you have 4 chem departments ? Physical, organic, inorganic, and theory ? (whatever) Not likely. As one person said, most CS programs have only 1 trivial course per term in first and second year, whose purpose is to get everyone up to speed by the time the real stuff comes around. This business about trivial programming courses is a crock. To go back to the chem example, programming is analogous to chem labs. If you can't do labs, you are not a chemist. If you can't program, you are not a computer scientist. There is simply no point in wasting 2 solid years on programming alone, just as there is no point in simply doing lab technique for a long time before theory is applied. >It is clear to me that the closest one needs to get to electrical >engineering to understand how computers execute software is the >level of combining NANDs and NORs in a logic design course. >Below that, the particular technology is irrelevant to computer science >since the logic is supposed to do the same thing no matter what the >technology is. Logic circuits may even be simulated on computers - so >who needs hardware anyway? :-). To some extent, this is true, although I would get more detailed about what a gate is, etc. However, my comment about CS programs being lax in the first 2 years still holds here: Frosh are doing nothing else, so they might as well learn electricity and a glossing over of transistors & what they do. CompEng is really the engineering school trying to offer CS. It shouldn't exist, simply because CS is too hardware-shy to begin with. >Of course it helps to know about your neighboring fields to understand >how your own field fits in. And it is necessary to have generalists who >understand more than two fields in order to integrate the neighborhoods. Specialization should be in 4th year. No earlier. >For proficiency requirements, I would like to see less focus on digital >electronics as well as numerical analysis. One hears this a lot from undergraduates, simply because they don't like numer- ical. I don't like it either (in general), but not teaching numerical is like not teaching quantum theory to a chemist. You may draw your own conclusions. >...but computer science should be out on its own by now. Numerical IS the cornerstone of CS. It (as Cleve Moler says) "is what God intended computers for". Live with it, it's fundamental. >I would like to see requirements that take into account the importance >of user interfaces (graphics, psychology of programming) and >application areas other than numerical analysis such as data bases This is plainly idiotic. It is inane to claim that enough is known about "soft" topics such as user interfaces to teach a 4-month undergraduate course on it. For CS to make it as a true academic discipline, it must have academic merit. In other words, it must be hard, and above all an undergraduate degree must cover "all" the bases after (say) third year. Currently the way it works is that programs cover all the good stuff in third year alone, while second year is left blank. In other disciplines, second year covers the general groundwork and the "history" of the field. Stuff like user interfaces are very much research topics and are hardly year 2 or 3 material. Certainly not cornerstones by any means. >Dan LaLiberte