root@bu-cs.UUCP (Barry Shein) (10/07/85)
Ok, fool that I am I will take a stab davidl@tekadg's complaints about UNIX. First off, the article does take on more of the tone of a flame and is filled with a barrage of insults to the UNIX community, though it does seem threaded deep into it are some things that might deserve response in spite of the ugly wrapper. I'll chalk it up to his frustration rather than deny him the right to be taken seriously. Second, his note seems to be very much concerned with the idea that the only people who would like UNIX only know UNIX, so, at the risk of total immodesty, I'll cite a little of my background to allay his suspicions (well, actually, I am not at all sure immodesty is the right term here): I have been in the computing business for over 10 years in several areas, including real-time, physiology, epidemiology, statistics, medical measurement, o/s, networking, compiler writing, AI, publishing a little etc. I have a Master's Degree in C.S. (not programming, C.S.!) and currently I am working on my PHD, have done a lot of programming on large IBM mainframes (BU currently owns two 3081s and almost all our stat stuff at Harvard [where I worked for several years] was done under CMS), TOPS-20, RT11, UNIX (most, name one), VMS and others. My position here at BU is Manager of Special Projects and I am a Lecturer in C.S. Perhaps this establishes a little credibility, there are certainly many people out there with a lot better credentials than mine, but your complaints about credentials seem to be limited to UNIX-only people with no CS background which I doubt will apply to me. Now, to try to respond to your note: 1. UNIX is not perfect, no one said it is. 2. Neither is anything else, the more you use a system the more flaws you find, this is not peculiar to UNIX. 3. In my experience I have found vendor supported O/S's to be horribly flawed, usually in their very specific view of what they believe people will be doing with their O/S and providing support only for that, often making it impossible or nearly so for a person with a problem outside of the vendor's original intentions (try dealing with files in general utilities on a system that supports a million file types, some such systems don't even have a useful 'copy' command, most have no utility which can convert from one file type to another *in general* [a few cases are usually possible] and even where they come close the amount of knowledge required of the user horrendous. Everyone complains about IBM/JCL, VMS library calls, UNIX command lines, so what else is new. There's no solution yet for various design trade-offs, unlike most other O/S's however, UNIX has and continues to go through massive re-designs in an attempt to improve the areas you cite while retaining the original design that seemed to be so appealing to almost everyone but you (even those wonderful native os's you speak of have been stealing like mad from UNIX, mostly because their user's are realizing that things like pipes and other UNIXisms are vastly superior ideas than what their systems offered, so they simulate, good for them I guess.) Unlike most O/S's I have seen, UNIX is one system that does not seem to have been designed primarly to beat out other systems on major govt RFP's (if you think most of those systems you know were designed for anything else, think again!) It was designed by people to get a myriad of projects done, mostly for in-house use, with a realization that people's time is at least as important as computer time. More importantly, when a bug or (particularly!) a design flaw is reported to these wonderful vendors o/s's you are almost always greeted with a big yawn, or at best it is passed by the marketing people to see if there is enough general interest to bother fixing the problem. In short, you are often stuck, with flawed software that you neither have the sources to nor any route to get fixed. It is easy for a UNIX site to obtain the sources. It ranges from inexpensive to moderate in price (about $43,000 for commercial sites.) Most importantly, that source distribution is real and it is workable. Most vendors that claim to offer sources for their system soon convince you that you really don't want them because they have made it unworkable to anyone outside of their organization. Try it if you don't believe me, just read the disclaimers on the source licenses they offer. Ask what that source license includes and how you get the rest (hint: it's probably not available.) This ability to control quality is often critical in an environment. The ability to fix security bugs/misfeatures that could put you out of the computing business alone is worth the price in most cases, let alone the ability to fix bugs as they occur rather than waiting for a vendor to get around to it (if they ever do.) In addition, the computing field is not a static one, the ability to experiment with new ideas, even within a fairly conservative organization, is often critical to progress, often this requires sources, even if just to share another site's progress before your vendor adopts it as a feature (if they ever will.) 4. UNIX's portability is not just a convenience, it is critical to an organization for the following reasons: a) It makes people portable. Even if you (as you claim) have had a lot of trouble learning UNIX, that trouble was well invested. I assume you have never used UNIX on an IBM mainframe, a CRAY-2 or other systems, but somehow I bet you will be much more productive right from the start rather than starting over again with whatever peculiar 'native' O/S they may offer. This may not be important to you personally, but it is important to management and not unreasonable. b) UNIX can (and does) follow hardware trends. Our graphics lab just switched from a VAX/750 to a Celerity, both running 4.2. You can barely tell the difference except that the Celerity runs 5-15X as fast as the 750 on the types of problems they are interested in, and it cost less than 1/2 what the 750 cost two years ago. They just wrote some tapes, unplugged the 750, plugged in the new machine, dumped the tapes and had things running a few days later. They were already sensitive to portability as they used the 3081 to run the same code in production, good for them, now they can do a fair amount of production on their 'clubhouse' machine. In contrast, there is a department here that just purchased an 8600 for about 10X what the celerity cost, and it is no faster even though they needed it for speed. Why? Because all their code they were running was tied into VMS and was not portable to anything but another VAX, what a waste! [actually, that's what they believed, who knows or cares really.] They could have each had a fully outfitted SUN or Celerity or some such as a PC on every desk for what they paid, now they get to fight for cycles. Sad, no? [note: the SUN3 will also, for about 1/10 the cost, give the 8600 a run for its money, the Celerity is a little more special purpose, the groups that bought Celereties are buying SUNs also, and it will all work together.] I am not saying there is no good reason to own a VAX (we own about 19 of them), just that that was a bad reason. I remember the VMS people here bragging about how much better the optimizer for Fortran is under VMS than UNIX, which is true, but WHAT GOOD does it do them now that they are stuck on vastly slower and/or less cost effective hardware? some of the machines I mentioned (and others) can run rings around their VAXes even with mediocre code optimizers and at much less cost. Funny how I don't hear from them these days. [Note: you specifically mentioned the ability of native O/S's to be more 'efficient' with the hardware than UNIX.] 5. Documentation - I could answer this faciley: I don't believe you, almost all computer documentation stinks. And when it tells me I can't do what I need to do (the case with most non-UNIX O/S's I have worked with) in simple terms, who cares. There are other sources of documentation, go to a good local University bookstore, I think you will see more books on how to use UNIX than all the other O/S's combined (maybe PC/DOS is close, but I don't think you were talking about PC's of that variety here.) Take one of the many UNIX training courses, if it's not worth it to you or your company to spend a few dollars, why should it be worth anything to the rest of us? (no disparagement of his company, that's a general statement.) 6) As far as USENIX conventions go, sorry, I have never been to one, but I find it hard to believe it is any different then the VMS groupies at a DECUS who go crazy when one of the VMS developers walks in. People are like that. To compare it to ACM or IEEE conferences is unfair, of course user group meetings are going to attract more people who are fans, why do you find this confusing? 7) The need for highly skilled professionals to manage complicated software environments is not limited to UNIX. We need it here for all our systems. The only difference seems to be that those vendor supplied systems never offer the software you need, so you go out and buy many random little third-party packages and spend a lot of time stopping them from fighting with the other random little third-party systems. And when an upgrade for the O/S comes little else gets done but trying to convince those random packages to run properly again, often for months. How many of you out there are running a word processing system from the same vendor as your vendor supplied system? Uh huh, what happened during the last (three times a year typically) upgrade to all that software? Got kinda broken, no? did the vendor care, of course not, did anything else get done by your 'gurus', of course not. I rarely find this happening on our UNIX systems, our UNIX people seem to spend all their time putting in new things, not fighting with vendors. Yeah, we do bug fixes to UNIX, but that's because we can. Your fantasy about running a computing environment without professionals is a common one. It happened here and a few departments ran off and bought vendor supplied O/S's, good for them. Now they come running to us because they find they have bought nothing of the sort, except that they have left the mainstream of progress here and are kinda stuck with what they bought. All they ever cry about is 'UNIX does it, make my system do it too [whatever 'it' is this week.] Look, it's like this, you could run a little medical care thing without real doctors as long as you are only getting toy problems, but once a real problem comes along, which always does, off you go running to the 'big-city' hospital for real professionals. The only way to avoid using specialists is to lower your standards (and as far as I'm concerned, feel free!) This is not peculiar to UNIX. I am sure a dentist can run a PC, so what has that got to do with running the networked system of a multi-million dollar organization with 40,000 users? Nothing, the latter takes professionalism and accountability. -Barry Shein, Boston University
ahby@meccts.UUCP (Shane P. McCarron) (10/11/85)
I also will make a fool of myself responding to the article by davidl@tekadg.UUCP: Facts: o I was first exposed to Unix just over a year ago. o I have a high school education o I have taken 1 CSci class at the University of Minnesota. o I am of approximately average computer programmer intelligence. I didn't have any trouble at all getting used to Unix. What the hell is your problem, Dave? Not only do you piss and moan about Unix, but you post your article separately to about 5 news groups that I subscribe to (just illustrating your vast misconceptions about news software, as well). Consequently I was forced to read your message 5 times... Note how happy this has made me. -- Shane P. McCarron Minnesota Educational Computing Corporation - Technical Services UUCP ihnp4!dicomed!meccts!ahby
makaren@alberta.UUCP (Darrell Makarenko) (10/13/85)
I can't help but respond to this one. > Facts: . . . > > o I am of approximately average computer programmer > intelligence. > > Shane P. McCarron > Minnesota Educational Computing Corporation - Technical Services What is the average intelligence of a computer programmer?
rfg@hound.UUCP (R.GRANTGES) (10/13/85)
[]
> What is the average intelligemce of a computer programmer?
Does anyone have any data? Let me guess: On this net, about IQ 132.
Remember,
--
"It's the thought, if any, that counts!" Dick Grantges hound!rfg
peterb@pbear.UUCP (10/15/85)
/* Written 9:01 pm Oct 12, 1985 by hound!rfg in pbear:net.general */ >[] >> What is the average intelligemce of a computer programmer? > >Does anyone have any data? Let me guess: On this net, about IQ 132. > >Remember, > >-- > >"It's the thought, if any, that counts!" Dick Grantges hound!rfg /* End of text from pbear:net.general */ Actually I always thought is was porportional to the time of day (in 24 hour format :-) ...
nather@utastro.UUCP (Ed Nather) (10/15/85)
> Consequently I was forced to read your message 5 > times... Note how happy this has made me. > Shane P. McCarron Quit crabbing and get your "n" key fixed. -- Ed Nather Astronomy Dept, U of Texas @ Austin {allegra,ihnp4}!{noao,ut-sally}!utastro!nather nather@astro.UTEXAS.EDU
mather@uicsl.UUCP (10/15/85)
Quote from "Uni-Wiz Weekly", vol 23, number 2: The `average' intelligence of a computer programmer can be broken down into 5 parts: 1) 4.2/5.0 B.S. gpa 2) An IQ from 90-110 3) Fluency in 3-4 computer languages, at least 1 human language 4) At least minimal knowledge of 3 Operating Systems 5) Near Comprehension of `recursion' ---- b.c.mather Software Surgeon uiucdcs!uicsl!mather
matt@prism.UUCP (10/16/85)
/* Written 9:01 pm Oct 12, 1985 by rfg@hound in prism:net.general */ > > What is the average intelligemce of a computer programmer? > > Does anyone have any data? Let me guess: On this net, about IQ 132. How about 127 -- fits in a signed char :-)
carl@aoa.UUCP (Carl Witthoft) (10/18/85)
In article <5300001@uicsl> mather@uicsl.UUCP writes: >The `average' intelligence of a computer programmer can be broken down >into 5 parts: > > 1) 4.2/5.0 B.S. gpa {and 4 more parts} Big deal! S/he got that gpa in computer science (major gut) (:==> ) Darwin's Dad (Carl Witthoft) ...!{decvax,linus,ima,ihnp4}!bbncca!aoa!carl @ Adaptive Optics Assoc., 54 Cambridgepark Dr. Cambridge, MA 02140 617-864-0201 " Buffet-Crampon R-13 , VanDoren B-45, and VanDoren Fortes ."