axness@teldata.UUCP (David Axness) (06/29/84)
I just returned from vacation to read this volume of news from various people telling me how much better VMS is than Unix. I cannot contain myself from responding to these messages. I worked for 2 1/2 years on a VMS system (11/780) where I was the system manager for the last year. I have been a Unix user for the last year. I have selected various lines from various articles to respond to. No personal attacks are intended. I respect the fact that we all have opinions so I am merely expressing mine. ===== W A R N I N G - Not for VMS lovers with weak stomachs ============= Remark #1 *** And VMS has EDT. This is not a bad editor so long as you promise never to use another terminal not blessed by DEC. It must also run at 9600 baud. Ever try to use it on an ADM3 or some other cheap terminal. Ever try to use it at 1200 baud. Good luck ! You also can't run system commands (like grep, sort, etc) from within the editor which is a standard practice with Unix editors. Remark #2 *** FORTRAN runs faster under VMS. "Real Programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies" - Taken from "Real Programmer Don't Write Specs" which appeared earlier in net.jokes. If your whole world revolves around FORTRAN, then you deserve VMS. Remark #3 *** VMS runs faster than Unix in a multiuser environment. What the Hell does that mean ! Lets take an example. I want to run a job to compile my code. It takes about 5 minutes so I want to run it in the background with my error messages going to a file. With Unix, I type the compile command as usual, append it with >& {filename} to redirect the output and finish it off with & so that it runs in the background. It sends a message to my terminal when its finished. With VMS, I have to go into that wonderful EDT editor and write a DCL command procedure that reassigns SYS$OUTPUT to a file and submits my compile to the proper batch queue. The average user will probably not be able to send a completion message back to the terminal although he/she could send a mail message. This whole mess shouldn't take more than 10 minutes assuming that the user is familiar with the 10 volume VMS manual set. I can now execute the command procdure to compile my job in the background. Not only is this wasted time, but it has distracted me from my intended purpose which is to develop software. ALSO, with the large number of software tools available on Unix, I can get more work done no matter how fast VMS can perform a context switch. ALSO, it is only a matter of time before fast machines designed to run Unix are readily available (have you seen a Pyramid 90X ?) for a lot less money that a DEC machine with VMS. Remark #4 *** One can always run a Unix emulator under VMS for *** those users who prefer Unix. I've used Eunice under VMS and it has a lot of problems. a. It is incredibly slow. b. It is not Unix - we could not port some C programs from Unix to Eunice. c. It is expensive. Remark #5 *** > No make. No awk....Perhaps there [are] DEC program *** products out there that make life bearable;... Yes, there are but do you know what they cost ! Thousands of $$ thats how much. And it doesn't stop there. The cost of software maintanence is also high. On the system I worked on, the cost of maintaining VMS and about 5 compilers was over $5,000 per year for a single VAX. GENERAL FLAME: If VMS is so great, get off the damn Unix net and get back to work on your wonderful VMS system. =============================================================================== I feel better now that I've said my peace. Have I left anything out Jeff ? Please feel free to respond negatively to this article but I probably won't read it beacause whatever it is, I've probably already heard it. Dave Axness
dyer@vaxuum.DEC (Creedence? I *love* Creedence...) (07/02/84)
Re: VMesS vs Unix Debate_______________________________________________________ Again, I have to correct a few misstatements. I'm presenting facts, not flames, so please take them with that in mind. Also, I'm presenting these as a knowledgable user of VMS, not as a DEC employee. >> And VMS has EDT. > This is not a bad editor so long as you promise never to use another terminal > not blessed by DEC. I'm no big fan of EDT, myself. I know it's extensible, though; if you don't have a VT52 or VT100 keypad, you can program the keys you do have. As for the screen handling, I believe you're right that EDT's limited. But a lot of terminals out there are VT100 clones, and it works just fine. > You also can't run system commands (like grep, sort, etc.) from within the > editor which is a standard practice with Unix editors. Unipress EMACS runs on VMS and can do all those things. There are a number of other editors that run on VMS (such as TECO) which are also very pow- erful. > I want to run a job to compile my code. It takes about 5 minutes so I want > to run it in the background with my error messages going to a file. > ... > With VMS, I have to go into that wonderful EDT editor and write a DCL command > procedure that reassigns SYS$OUTPUT to a file and submits my compile to the > proper batch queue. The average user will probably not be able to send a > completion message back to the terminal although he/she could send a mail > message. This whole mess shouldn't take more than 10 minutes assuming that > the user is familiar with the 10 volume VMS manual set. I can now execute > the command procedure to compile my job in the background. Not only is this > wasted time, but it has distracted me from my intended purpose which is to > develop software. Instead of familiarizing yourself with the VMS manual set, you should have used online HELP. There you could easily see that $ SPAWN/NOWAIT/OUTPUT=file your-command would do exactly what you want. This doesn't notify you when it's done, but one can work around that very easily. Or you could run it in batch. This requires a command file, yes, but unless you write perfect code every time, you'll use that file more than once. And it doesn't take long to write a three-line command file, regardless of what editor you use. Submitting it is as simple as $ SUBMIT/LOG_FILE=file/NOTIFY command-file The /LOG_FILE puts your error messages in a file. The /NOTIFY tells you when it's done. Finally, all VMS compilers have a /LIST=file switch that puts error messages in a file along with the source. Now, you may prefer tokens like `>&' and such over qualifiers like /OUTPUT; I agree, somewhat: they save typing. They *are* cryptic, though, and the naive user might find English words easier. Anyways, frequently-typed com- mands can be replaced with command symbols. >> One can always run a Unix emulator under VMS for users who prefer Unix. > I've used Eunice under VMS and it has a lot of problems. DEC/Shell works very well. It makes a lot of Unix utilities available to VMS users. (Then again, many of those utilities were already made available with DECUS C, as mentioned before.) And as a spinoff, VMS can now use pipes. <_Jym_> :::::::::::::::: Jym Dyer ::::' :: `:::: Nashua, New Hampshire ::' :: `:: :: :: :: DYER%VAXUUM.DEC@DECWRL.ARPA :: .::::. :: {allegra|decvax|ihnp4|ucbvax}!decwrl!dec-rhea!dec-vaxuum!dyer ::..:' :: `:..:: ::::. :: .:::: Statements made in this article are my own; they might not :::::::::::::::: reflect the views of |d|i|g|i|t|a|l| Equipment Corporation.
steve@ea.UUCP (07/05/84)
#R:teldata:-41100:ea:13400007:000:4035 ea!steve Jul 4 18:49:00 1984 /***** ea:net.unix / teldata!axness / 2:37 am Jun 30, 1984 */ I just returned from vacation to read this volume of news from various people telling me how much better VMS is than Unix. I cannot contain myself from responding to these messages. I worked for 2 1/2 years on a VMS system (11/780) where I was the system manager for the last year. I have been a Unix user for the last year. I have selected various lines from various articles to respond to. No personal attacks are intended. I respect the fact that we all have opinions so I am merely expressing mine. ===== W A R N I N G - Not for VMS lovers with weak stomachs ============= Remark #1 *** And VMS has EDT. This is not a bad editor so long as you promise never to use another terminal not blessed by DEC. It must also run at 9600 baud. Ever try to use it on an ADM3 or some other cheap terminal. Ever try to use it at 1200 baud. Good luck ! You also can't run system commands (like grep, sort, etc) from within the editor which is a standard practice with Unix editors. Remark #2 *** FORTRAN runs faster under VMS. "Real Programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies" - Taken from "Real Programmer Don't Write Specs" which appeared earlier in net.jokes. If your whole world revolves around FORTRAN, then you deserve VMS. Remark #3 *** VMS runs faster than Unix in a multiuser environment. What the Hell does that mean ! Lets take an example. I want to run a job to compile my code. It takes about 5 minutes so I want to run it in the background with my error messages going to a file. With Unix, I type the compile command as usual, append it with >& {filename} to redirect the output and finish it off with & so that it runs in the background. It sends a message to my terminal when its finished. With VMS, I have to go into that wonderful EDT editor and write a DCL command procedure that reassigns SYS$OUTPUT to a file and submits my compile to the proper batch queue. The average user will probably not be able to send a completion message back to the terminal although he/she could send a mail message. This whole mess shouldn't take more than 10 minutes assuming that the user is familiar with the 10 volume VMS manual set. I can now execute the command procdure to compile my job in the background. Not only is this wasted time, but it has distracted me from my intended purpose which is to develop software. ALSO, with the large number of software tools available on Unix, I can get more work done no matter how fast VMS can perform a context switch. ALSO, it is only a matter of time before fast machines designed to run Unix are readily available (have you seen a Pyramid 90X ?) for a lot less money that a DEC machine with VMS. Remark #4 *** One can always run a Unix emulator under VMS for *** those users who prefer Unix. I've used Eunice under VMS and it has a lot of problems. a. It is incredibly slow. b. It is not Unix - we could not port some C programs from Unix to Eunice. c. It is expensive. Remark #5 *** > No make. No awk....Perhaps there [are] DEC program *** products out there that make life bearable;... Yes, there are but do you know what they cost ! Thousands of $$ thats how much. And it doesn't stop there. The cost of software maintanence is also high. On the system I worked on, the cost of maintaining VMS and about 5 compilers was over $5,000 per year for a single VAX. GENERAL FLAME: If VMS is so great, get off the damn Unix net and get back to work on your wonderful VMS system. =============================================================================== I feel better now that I've said my peace. Have I left anything out Jeff ? Please feel free to respond negatively to this article but I probably won't read it beacause whatever it is, I've probably already heard it. Dave Axness /* ---------- */
henry@utzoo.UUCP (Henry Spencer) (07/05/84)
Jym Dyer says: Now, you may prefer tokens like `>&' and such over qualifiers like /OUTPUT; I agree, somewhat: they save typing. They *are* cryptic, though, and the naive user might find English words easier. Anyways, frequently-typed com- mands can be replaced with command symbols. It is a popular misconception, almost totally without foundation, that naive users find an English word easier to use than an arbitrary symbol, when both are being used as "magic words" which must be exactly right. The small amount of real evidence that exists suggests that, in the presence of a help facility at least, the naive users actually find cryptic symbols *easier* to use correctly. The reasons are somewhat conjectural, but the obvious one is that an arbitrary bit of gibberish does not carry the mass of extra connotations that an English word does, and hence is less likely to be used incorrectly in the belief that "the computer understands English". Yes, Virginia, people make fewer errors if you call it "grep" than if you call it "search". Really. Check out the papers in the 1983 ACM SIGCHI proceedings if you doubt me -- I'd give a more exact reference, but I don't have the proceedings handy to pick out the particular paper. "`User-friendly' really only means marketing-friendly." -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,linus,decvax}!utzoo!henry
gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) (07/08/84)
In support of Henry's point, note that mathematical notation consists primarily of small symbols (a few abbreviations of words are used in category theory). Only for a totally na"ive user would it be appropriate to require communication of concepts via natural-language words. As has been observed before, the Bourne shell (or equivalent) is not an appropriate interface for the "man in the street", but it works well for trained operators.
SMH@SRI-KL.ARPA (07/08/84)
From: "Scott M. Hinnrichs" <SMH@SRI-KL.ARPA> RE: "User-Friendly" systems People are always looking for reasons to keep from having to learn new things (read: "do work"). "User-Friendliness" is just another bastion of defense against having to do work. Another free lunch myth. My favorite ploy to get someone past this barrier and onto learning, and eventually doing, is to get them to realize this themselves. One comment I get occasionally from people is that in order to use UNIX they really have to know what they are doing. Wow. That is bad enough, but I go one step further and ask them about the system they are currently using: Q. "Didn't it take time to learn the system you are using now?" A. "Well, yes, but I don't know much about that system either." "User-Friendliness" <--> "User-Laziness" The fact of the matter is that anything worth doing is going to take time and pain, and most people are not willing to go to the lengths necessary to attain their goals. When someone is confronted with the hurdle of learning how to work with a new tool it helps to address their cries of "User-Unfriendliness" by helping them to see the overall picture. The greatest hurdle is their fear of the nebulous unknown. In-other-words you have to kill the boogie-man created by all the wonderful hearsay spread about computers, UNIX, etc. Once you get them through all their preliminary fears they can start learning. The time spent learning is really quite minimal. The real time sink is convincing them to learn. It is quite rewarding to see someone who was incredibly biased against what you are teaching cut through the BS and see things clearly. They even seem to have fun, after-a-fashion. As far as "User-Friendliness" is concerned they tend to spout that phrase less and less as time goes by. However, you still hear it from them occasionally; whenever they don't understand what they are doing. These comments are as much for me as they are for you. I am about to start teaching another round of UNIX classes here at SRI and I am convincing myself it's worth the effort to take the time to teach, just the same as I have to convince them it's worth the effort to take the time to learn. Here's to a "User-Friendly" life! Scott M. Hinnrichs SRI International sri-unix!smh, smh@sri-unix -------
hopp@nbs-amrf.UUCP (Ted Hopp) (08/17/84)
From "Scott M. Hinnrichs" <SMH@SRI-KL.ARPA> we get: > RE: "User-Friendly" systems > People are always looking for reasons to keep from having to > learn new things (read: "do work"). "User-Friendliness" is just another > bastion of defense against having to do work. Another free lunch myth. > > ... > > "User-Friendliness" <--> "User-Laziness" > The fact of the matter is that anything worth doing is going to > take time and pain, and most people are not willing to go to the lengths > necessary to attain their goals. Just a few random thoughts this provokes: Reminds me of when I took up snow skiing. The motto then (late 60's) concerning ski boots was, "If they don't hurt, they don't fit right." One of the fundamentals of cult psychology is that the higher the dues, the more the member values the membership. This is independent of the inherent value of membership. Does anyone know why we dropped the nice, time-consuming, and painful Roman number system for the user-friendly Arabic number system, other than pure laziness? Until someone convinces me that there ain't no better way, I will continue to consider all the bad-mouthing of user-friendliness just so much garbage. -- Ted Hopp UUCP: {seismo,allegra}!umcp-cs!nbs-amrf!hopp National Bureau of Standards ARPA: hopp.nbs-amrf.umcp-cs@udel-relay Metrology A127 BELL: (301)921-2461 Gaithersburg, MD 20899