jdudeck@polyslo.CalPoly.EDU (John R. Dudeck) (02/26/90)
>The problem with OS/2 is that it is simply Unix reimplemented. And as Henry >Spencer point out, "those who don't understand Unix are doomed to reinvent >it, poorly". I like that. In fact the discussion of operating systems really boils down to a comparison of OS/2 and Unix. I think it is worthwhile to do so (but maybe not on this newsgroup). >Now, if you were a software developer, and you had the resources to make >your program work under Unix or OS/2, but not both, which would you choose? >An OS that runs only on 286es, or an OS that runs on 286s AND everything else >imaginable (and a few that aren't). The answer to this is not as clear-cut as you might think. It all depends on which platform your customer base is on. It depends on what type of application it is. If it is GUI-based, then you have even more variables to consider. OS/2 is not just for 286's. It isn't 386-specific, if that is what you mean. While Unix is available on practically all hardware platforms, there aren't many Unix boxes sitting on people's desks. But there are a lot of potentially OS/2 boxes on people's desks. If you want to sell a program from Egghead Software for Unix, you've either got to sell source code in c, or else sell binaries for each different flavor of hardware and Unix out there. And then ask Joe User to compile your program on his Unix system. On on OS/2 system he would have just one version to select from at the store, and he would have it installed and running with a lot less trouble. Or to put it another way, if you were Joe User, and had a choice between Unix and OS/2 which would you buy? I'm sorry to keep this discussion going on this group. I shall refrain from further comments. -- John Dudeck "You want to read the code closely..." jdudeck@Polyslo.CalPoly.Edu -- C. Staley, in OS course, teaching ESL: 62013975 Tel: 805-545-9549 Tanenbaum's MINIX operating system.
AMillar@cup.portal.com (Alan DI Millar) (03/13/90)
>One of the problems with Unix on a PC is this: > > Who on Earth needs a multi-user PC? > >A multi-user NeXT, I can understand, but a PC? Why? What makes a 386 in IBM pc hardware any different that a 386 in a Sun 386i? And apparently you haven't seen any of the 486 machines yet :-) - Alan Millar
phil@pepsi.amd.com (Phil Ngai) (03/15/90)
In article <7989@sdcsvax.UCSD.Edu> velasco@beowulf.UCSD.EDU (Gabriel Velasco) writes: |It's true that one of the things that has made UNIX so popular is the way that |it handles multiple users, but this is not the only thing. History mechanisms, Be serious. I use both DOS and Unix daily and I have a much better history mechanism on DOS than I do on Unix. |job control with backgrounding, programming interface, are just a few more I have job control and backgrounding too. You don't HAVE to have Unix to do this. -- Phil Ngai, phil@amd.com {uunet,decwrl,ucbvax}!amdcad!phil Boycott the census! With the history of abuse census data has, do you really trust the government?
mollers.pad@nixpbe.uucp (Moellers) (03/15/90)
CMH117@psuvm.psu.edu (Charles Hannum) writes: >One of the problems with Unix on a PC is this: > Who on Earth needs a multi-user PC? >A multi-user NeXT, I can understand, but a PC? Why? I agree! Who needs PCs anyhow? They're ancient, their architectural evolution is blocked by memories of days gone by. Any development of PCs and their software will have to be compatible with 10 64K segments, irrespectable of whether the CPU is an 8086 with (max.) 16 Segments of 64KB each or an 80386 with n segments of 4 GB each! (I have heard the rumor that Microsoft was forced by IBM to reduce the functionality of MS-Windows in order for it to be runnable under MESSDOS. But, let's face it: - There is a LOT OF software (especially games) out there running under MESSDOS. - There are MESSDOS-emulators (VP/ix, MERGE-368) that allow You to run this SW under UNIX. Personally, I would prefer an NS32k based system. That CPU is fully upward and downward compatible within it's family! Ce la vie! Josef Moellers | c/o Nixdorf Computer AG USA: uunet!philabs!linus!nixbur!mollers.pad | Abt. DX-PC !USA: mcvax!unido!nixpbe!mollers.pad | Pontanusstr Phone: (+49) 5251 146245 | D-4790 Paderborn
srg@cunixd.cc.columbia.edu (Steven R Gerber) (03/15/90)
In article <29499@amdcad.AMD.COM> phil@pepsi.AMD.COM (Phil Ngai) writes: >In article <7989@sdcsvax.UCSD.Edu> velasco@beowulf.UCSD.EDU (Gabriel Velasco) writes: >|It's true that one of the things that has made UNIX so popular is the way that >|it handles multiple users, but this is not the only thing. History mechanisms, > >Be serious. I use both DOS and Unix daily and I have a much better >history mechanism on DOS than I do on Unix. Did you mean you're using an additional utility (4DOS, stack, etc.)? If you mean the regular F3 etc., you gotta be kidding! > >|job control with backgrounding, programming interface, are just a few more > >I have job control and backgrounding too. You don't HAVE to have Unix >to do this. True. You can kludge this (Windows, DesqView, etc.). But FACE IT: MS-DOS just plain ain't got it. I don't want to do back-flips to code. BTW, have you every programmed for Windows? It's a real pain-in-the-ass and the SDK is expensive. > > >-- >Phil Ngai, phil@amd.com {uunet,decwrl,ucbvax}!amdcad!phil >Boycott the census! With the history of abuse census data has, >do you really trust the government? **************************************************************** * Steven R. Gerber - PAL (Programmer At Large) * srg@cunixd.cc.columbia.edu * Tel: 212-794-8721 * UUCP: ...rutgers!columbia!cunixd!srg * FAX: 212-794-8721 ****************************************************************
schaut@cat9.cs.wisc.edu (Rick Schaut) (03/16/90)
What about multi-processing (not multi-tasking which both systems can do, and not multi-user which is the UNIX flavor)? How do the two systems compare in a multi-processor environment? UNIX is exclusively designed to work with losely coupled CPU's (i.e. it's pretty good for a network) but you can't put it on a system with tightly coupled CPU's without adding some extensions (ask the people at Cray). OS/2, on the other hand, has the ability to spin off multiple threads; exactly the feature one needs to take advantage of a couple of different CPU's which have access to the same memory. Also, in conjunction with the Lan Manager, OS/2 gets the ability to work well with loosely coupled CPU's as well. So, in the ability to handle future hardware, score: OS/2 1, UNIX 1/2 In a world of multi-processor workstations and distributed processing facilities (where multi-user capabilties aren't necessary) it looks as though OS/2 would be the operating system of choice. -- Rick (schaut@garfield.cs.wisc.edu) "I'm a theory geek; we use Turing machines!"--Gary Lewandowski
schaut@cat9.cs.wisc.edu (Rick Schaut) (03/16/90)
In article <90070.224516DLV101@psuvm.psu.edu> DLV101@psuvm.psu.edu (Dwaine VanBibber) writes: | A lot of people need a multi-user PC!! Well, give me a for-instance. One interesting development within the department here at UW is that large multi-user machines are being replaced by networked workstations with distributed processing facilities. Where in such an environment do I need multi-user capabilities? | I agree with one of the previous opinions that OS/2 is just Unix reinvented. | Its not really necessary. You can run existing DOS applications on a Unix | box..more than one at a time also..which you can't do with OS/2. Show me a _standard_ version of UNIX that can handle tightly coupled CPU's in a multi-processor workstation. OS/2 is most definitely _not_ UNIX reinvented. | They should have just developed a REALLY GOOD GUI (PM or even NeXTStep) to run | ontop of Unix. And they ought to ship the damned development tools with it, | as Sun does with Openwindows. If I have 30 workstations for 20 users and 10 developers, you would have me paying for the "damned development tools" for the 20 workstations being used by people who aren't doing any development. It's a good thing you don't run a business. You'd run it into the ground. -- Rick (schaut@garfield.cs.wisc.edu) "I'm a theory geek; we use Turing machines!"--Gary Lewandowski
malloy@nprdc.arpa (Sean Malloy) (03/16/90)
In article <4473@daffy.cs.wisc.edu> schaut@cat9.cs.wisc.edu (Rick Schaut) writes: >UNIX is exclusively designed to work with losely coupled CPU's (i.e. it's >pretty good for a network) but you can't put it on a system with tightly >coupled CPU's without adding some extensions (ask the people at Cray). >OS/2, on the other hand, has the ability to spin off multiple threads; >exactly the feature one needs to take advantage of a couple of different >CPU's which have access to the same memory. Also, in conjunction with >the Lan Manager, OS/2 gets the ability to work well with loosely coupled >CPU's as well. >So, in the ability to handle future hardware, score: OS/2 1, UNIX 1/2 >In a world of multi-processor workstations and distributed processing >facilities (where multi-user capabilties aren't necessary) it looks as >though OS/2 would be the operating system of choice. Unix gets 1/2 because it works well with loose-coupled CPUs but needs an extension to work with tightly-coupled CPUs. OS/2 gets 1 because it works well with tightly-coupled CPUs but needs an extension to work with loosely-coupled CPUs. Is there some kind of bias here you're not telling us about? I would give each of them 1/2 for only handling one of the configurations, and then hand OS/2 a bit more for having LAN Manager more conveniently available. Then I'd take a bunch back because OS/2 is designed for a single processor architecture, and some more because it doesn't handle it right (only one 'DOS compatibility box' thread running at one time, for example). When OS/2 runs on something other than 80[234]86 machines, it's going to be a serious competitor for future operating systems; as long as it's architecture-specific, regardless of how well it runs on that architecture, it's still a dodo. Sean Malloy | "The Crystal Wind is the Navy Personnel Research & Development Center | Storm, and the Storm is Data, San Diego, CA 92152-6800 | and the Data is Life." malloy@nprdc.navy.mil | -- _Emerald Eyes_, D.K. Moran
phil@pepsi.amd.com (Phil Ngai) (03/16/90)
In article <1990Mar15.152746.21025@cunixf.cc.columbia.edu> srg@cunixd.cc.columbia.edu (Steven R Gerber) writes: |>Be serious. I use both DOS and Unix daily and I have a much better |>history mechanism on DOS than I do on Unix. | Did you mean you're using an additional utility (4DOS, stack, etc.)? | If you mean the regular F3 etc., you gotta be kidding! Yes, I have some additional utilities on DOS. I do on Unix too, things like readnews, emacs, ksh, etc. |>I have job control and backgrounding too. You don't HAVE to have Unix |>to do this. | True. You can kludge this (Windows, DesqView, etc.). | But FACE IT: MS-DOS just plain ain't got it. I refuse to accept comparisons of DOS without any utilities unless you also look at Unix without utilities like csh, X, etc. | I don't want to do back-flips to code. BTW, have you every programmed | for Windows? It's a real pain-in-the-ass and the SDK is expensive. Have you ever programmed for X? Are you saying Unix is not expensive? -- Phil Ngai, phil@amd.com {uunet,decwrl,ucbvax}!amdcad!phil Boycott the census! With the history of abuse census data has, do you really trust the government?
pml@cacs.usl.edu (Patrick M. Landry) (03/16/90)
In article <4475@daffy.cs.wisc.edu> schaut@cat9.cs.wisc.edu (Rick Schaut) writes: >In article <90070.224516DLV101@psuvm.psu.edu> DLV101@psuvm.psu.edu (Dwaine VanBibber) writes: >| A lot of people need a multi-user PC!! >Well, give me a for-instance. >One interesting development within the department here at UW is that large >multi-user machines are being replaced by networked workstations with >distributed processing facilities. Where in such an environment do I need >multi-user capabilities? > You need the concept of multiple users as soon as you want to share things. Access permissions kinda rely on the OS having the notion of "other users". This goes for sharing the machines, files, etc. This is one of the biggest problems with networking a bunch of DOS boxes together; its damned hard to share things even though that's what the net is presumably for. >-- >Rick (schaut@garfield.cs.wisc.edu) > >"I'm a theory geek; we use Turing machines!"--Gary Lewandowski pad pad pad pad pad -- patrick pml@cacs.usl.edu
wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) (03/16/90)
In article <6508@skinner.nprdc.arpa> malloy@nprdc.arpa (Sean Malloy) writes: >....... >I would give each of them 1/2 for only handling one of the >configurations, and then hand OS/2 a bit more for having LAN Manager >more conveniently available. Then I'd take a bunch back because OS/2 is >designed for a single processor architecture, and some more because it >doesn't handle it right (only one 'DOS compatibility box' thread >running at one time, for example). > >When OS/2 runs on something other than 80[234]86 machines, it's going >to be a serious competitor for future operating systems; as long as >it's architecture-specific, regardless of how well it runs on that >architecture, it's still a dodo. > > > Sean Malloy | "The Crystal Wind is the > Navy Personnel Research & Development Center | Storm, and the Storm is Data, > San Diego, CA 92152-6800 | and the Data is Life." > malloy@nprdc.navy.mil | -- _Emerald Eyes_, D.K. Moran OS/2 ver 2.0 is supposed to handle multiple 'DOS compatibility boxes' and with a lot more safety due to it use of the Virtual Machine modes provided by the 80386. Of course it will require a 80386. Another thing, check out the latest version OS/2 with LAN Manager from Compaq (sp?) that is written for the System Pro (tm). It is designed to use two 80386's or 80486's when the cards come out this summer. There are reviews in the latest PC Magazine and Byte magazines about the System Pro and they talk a little about the LAN Manager written for it. One last thing, rumor has it that Bill Gates is trying to get OS/2 rewritten in C so that it can be easily transported to new machines. Dean Wallwey
velasco@beowulf.ucsd.edu (Gabriel Velasco) (03/16/90)
schaut@cat9.cs.wisc.edu (Rick Schaut) writes: >Well, give me a for-instance. >One interesting development within the department here at UW is that large >multi-user machines are being replaced by networked workstations with >distributed processing facilities. Where in such an environment do I need >multi-user capabilities? In a "truely" distributed system, when you logged onto your workstation, you would not know which workstation's cpu was taking care of you. Your jobs would be loaded onto whichever workstation had the lightest load. The whole network would act as a single, multiprocessor, multitasking, multiuser system. This is one of the goals of distributed systems research. ________________________________________________ <>___, / / | ... and he called out and said, "Gabriel, give | /___/ __ / _ __ ' _ / | this man an understanding of the vision." | /\__/\(_/\/__)\/ (_/_(/_/|_ |_______________________________________Dan_8:16_|
les@chinet.chi.il.us (Leslie Mikesell) (03/16/90)
In article <18459@boulder.Colorado.EDU> wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) writes: >>I would give each of them 1/2 for only handling one of the >>configurations, and then hand OS/2 a bit more for having LAN Manager >>more conveniently available. AT&T is well on the way (if it isn't already out) to having an OS/2 LAN Manager compatible server running under unix - with extensions for things like spooled printing back to the client PC's which their older starlan DOS server provided. Being unix based it will (a) run on CPU's other than Intel's, and (b) maintain a concept of file ownership for files stored on the server. I'm not familiar with LAN manager - can you make certain files writable by one user, readable by others? Will it maintain the attributes if you back the files up and restore them somewhere else? >One last thing, rumor has it that Bill Gates is trying to get OS/2 >rewritten in C so that it can be easily transported to new machines. Gee, I wonder where he got that idea? Maybe he will add the user and group id concept as well. Les Mikesell les@chinet.chi.il.us
lws@comm.WANG.COM (Lyle Seaman) (03/17/90)
CMH117@psuvm.psu.edu (Charles Hannum) writes: >One of the problems with Unix on a PC is this: > Who on Earth needs a multi-user PC? >A multi-user NeXT, I can understand, but a PC? Why? So I can play xtank with all the members of my family, of course. On a related note(?), the term PC was originally coined to refer to a basically single-user machine, that just -happened- to have a microprocessor CPU. Now that PCs are going multi-user, when are we going to dump the PC moniker, and start referring to these as simply microcomputers (slightly more accurate)? Followups to ?? -- Lyle sendmail.cf under construction, pardon the From: lws@comm.wang.com (or, uunet!comm.wang.com!lws) (508) 967-2322
lws@comm.WANG.COM (Lyle Seaman) (03/17/90)
wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) writes: >That's the whole reason for OS/2...To be a true multitasking operating >system, with out the overhead of multiuser! Actually, multiuser abilities don't add a whole lot of overhead to the actual OS, though they do add a couple of necessary ancillary programs. Multitasking is far costlier. >I agree, multitasking is needed even though not obviously to the user. >A good example of this is even a spread sheet. It can calculate in the >background, while printing out a sheet, while the user is moving around >the screen all at the same time! I'm not sure this is such a great example. Any spreadsheet worth its salt would do this on a singlethreaded OS. This is all one application and isn't terribly difficult to do. A better example is the ability to do away with TSRs and other klduges and write _real_ communicating processes without a global shared address space. >> >> Actually, to sidetrack a bit - how well does OS/2 really multi-process? >>Is it capable of keeping an accurate file transfer running in the background >>using one serial port while another serial/parallel port is also in use? >OS/2 can do a lot more than that if you have a machine with enough >memory and speed. To give an example... OS/2 can be downloading stock >information into Excel (a super power spread sheet!), which updates one >of its graph, which happens to be "hotlinked" into a PageMaker Document you >are working on. At the same time, a large C program can be compiling >in the back ground, while another application is printing out. >ALL AT THE SAME TIME!!!!!!!(if you have a fast system with LOTS of >memory) These examples are not nearly as useful as the original proposed example. So I repeat: Is it capable of keeping an accurate file transfer running in the background using one serial port while another serial port is also in use? >>If >>not, then OS/2 is little more than another crippled memory-hog operating >>system - something to avoid. It is too bad that a MS-DOS compatible machine >>cannot be designed from the ground up to multitask - I had hoped that 386 >>machines would be capable of it. >Don't confuse the operating system with machine! RIGHT. >286 machines and more importantly, 386 machines were designed for >multitasking! RIGHT. > MS-DOS wasn't!!!!!!!!! Thus OS/2 was created!!!! WRONG. I can't believe that OS/2 was created. It shakes my faith in a benevolent God. OS/2 must simply have happened, one of those inexplicable mysteries that plagues mankind. Oh well, maybe I can blame IBM... Besides, OS/2 is just warmed-over MS-DOS. I'd rather have TRS-DOS 4.0. -- Lyle sendmail.cf under construction, pardon the From: lws@comm.wang.com (or, uunet!comm.wang.com!lws) (508) 967-2322
wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) (03/17/90)
In article <1990Mar16.155157.5431@chinet.chi.il.us> les@chinet.chi.il.us (Leslie Mikesell) writes: >...... >>One last thing, rumor has it that Bill Gates is trying to get OS/2 >>rewritten in C so that it can be easily transported to new machines. > >Gee, I wonder where he got that idea? Maybe he will add the user >and group id concept as well. > >Les Mikesell > les@chinet.chi.il.us The new file system (HPFS) that comes with OS/2 is supposed to be a lot more modifiable than that of *NIX. It's supposed to be alot more "Object oriented, with editable attributes". So if you personally wanted to put user and goup id's on the files go ahead, if you want to put ICON symbols or anything else in the attributes, go ahead! OS/2's file system will actually be alot more versatile than *NIX's! By the way, I think these are some of the features that LAN Manager (tm) uses to make a lot more coherent network. Dean Wallwey
GILLA@QUCDN.QueensU.CA (Arnold G. Gill) (03/17/90)
In article <1990Mar16.222205.9749@comm.WANG.COM>, lws@comm.WANG.COM (Lyle Seaman) says: > >wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) writes: > >>>If >>>not, then OS/2 is little more than another crippled memory-hog operating >>>system - something to avoid. It is too bad that a MS-DOS compatible machine >>>cannot be designed from the ground up to multitask - I had hoped that 386 >>>machines would be capable of it. > >>Don't confuse the operating system with machine! > >RIGHT. > >>286 machines and more importantly, 386 machines were designed for >>multitasking! > >RIGHT. Sorry, I just can't buy that at all. x86 chips may be designed for multitasking, but the AT style machines that are being sold are not. You really want to redesign the machine from the ground up to be multitasking, and these are not. If you want to look at real multitasking PCs, the ones to consider are the Amiga and possibly the Mac II. It is likely that the Amiga is still the best machine around, but because it doesn't have the Big Blue label on it, and is only MS-DOS compatible after an extra card is added, it doesn't have the exposure the IBM-PC family has. ------- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- | Arnold Gill | | | Queen's University at Kingston | If I hadn't wanted it heard, | | BITNET : gilla@qucdn | I wouldn't have said it. | | X-400 : Arnold.Gill@QueensU.CA | | | INTERNET : gilla@qucdn.queensu.ca | | -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
kdq@demott.COM (Kevin D. Quitt) (03/18/90)
In article <27830@cup.portal.com> AMillar@cup.portal.com (Alan DI Millar) writes: >... >What makes a 386 in IBM pc hardware any different that a 386 in a Sun 386i? For one thing, it only has half the thruput of a Sun 386i. kdq -- Kevin D. Quitt Manager, Software Development DeMott Electronics Co. VOICE (818) 988-4975 14707 Keswick St. FAX (818) 997-1190 Van Nuys, CA 91405-1266 MODEM (818) 997-4496 Telebit PEP last 34 12 N 118 27 W srhqla!demott!kdq kdq@demott.com "Next time, Jack, write a God-damned memo!" - Jack Ryan - Hunt for Red Oct.
jdudeck@polyslo.CalPoly.EDU (John R. Dudeck) (03/18/90)
In article <1990Mar16.222205.9749@comm.WANG.COM> lws@comm.WANG.COM (Lyle Seaman) writes: >> MS-DOS wasn't!!!!!!!!! Thus OS/2 was created!!!! > >I can't believe that OS/2 was created. It shakes my faith >in a benevolent God. OS/2 must simply have happened, one of those >inexplicable mysteries that plagues mankind. Oh well, maybe I can >blame IBM... > >Besides, OS/2 is just warmed-over MS-DOS. I'd rather have TRS-DOS 4.0. I am really enjoying this discussion, because I have great hopes for OS/2's future. It *sounds like* everything I want in an OS for my 386. But, on the other hand, I was reminded recently by another netter of Frederick Brooks' wisdom in the book, The Mythincal Man Month. He has a chapter in it entitled, "Self-Discipline--The Second-System Effect". I quote: "An architect's first work is apt to be spare and clean. He knows he doesn't know what he is doing, so he does it carefully and with great restraint. "As he designs the first work, frill after frill and embellishment after embellishment occur to him. These get stored away to be used "next time". Sooner or later the first system is finished, and the architect, with firm confidence and a demonstrated mastery of that class of systems is ready to build a second system. "This second system is the most dangerous system a man ever designs. When he does his third and later ones, his prior experiences will confirm each other... "The general tendency is to over-design the second system, using all the ideas and frills that were cautiously sidetracked on the first one. The result, as Ovid says, is a "big pile"." Just looking at OS/2 it is hard to not think of it as being somebody's "second system". Still, I don't think that it is all that bad. It's something with tremendous potential, but the complexity is making it very difficult to develop and really finish. On the other hand, Unix is big and complex, too, and it has taken a long time to get where it is. But it started out simple. The difference is that Unix started off with a good design that has lent itself to incremental evolutionary improvements without destroying the design. MS-DOS was a very poor place to start, and such add-ons as Windows will never make it into a better system. Hence the impetus for OS/2. OS/2 is not an evolutionary change to MS-DOS. It is a completely new system, with a "DOS compatibility" appendage. What amazes me is that what most users seem to want from OS/2 is the ability to use it as a multitasing DOS system. Everyone seems to be complaining that the current version only has *one* compatibility box. The value of OS/2 is not in its compatibility box! -- John Dudeck "You want to read the code closely..." jdudeck@Polyslo.CalPoly.Edu -- C. Staley, in OS course, teaching ESL: 62013975 Tel: 805-545-9549 Tanenbaum's MINIX operating system.
davidsen@sixhub.UUCP (Wm E. Davidsen Jr) (03/18/90)
{ note: I have put followup to comp.unix.questions, since this group has nothing to do with unix OR os/2. I have sent many people notes asking them to move the thread, hopefully people will cooperate. In article <4473@daffy.cs.wisc.edu> schaut@cat9.cs.wisc.edu (Rick Schaut) writes: | UNIX is exclusively designed to work with losely coupled CPU's (i.e. it's | pretty good for a network) but you can't put it on a system with tightly | coupled CPU's without adding some extensions (ask the people at Cray). Except for Solbourne running multicpu SunOS, someone (concetric?) running multiprocessor SCO Xenix on the Zenith Z1000, Encore, Alliant, Convex and Cray, etc. In short, your statement just isn't true. | So, in the ability to handle future hardware, score: OS/2 1, UNIX 1/2 | In a world of multi-processor workstations and distributed processing | facilities (where multi-user capabilties aren't necessary) it looks as | though OS/2 would be the operating system of choice. OS/2 has threads. Many variants of UNIX do, too, although there isn't a standard. UNIX allows multiple users, not only at once, but allows multiple people to share a machine without depending solely on their being willing and able to avoid causing each other problems. OS/2 doesn't have the concept of file ownership and protection, and currently anyone can change process priorities. OS/2 is easily postable to any Intel 80X86 CPU, unix is portable to any CPU with memory management (and some without, more or less). OS/2 gives you your choice of vendors of utilities, UNIX comes with utilities. OS/2 is the operating system of Intel and IBM's choice. -- bill davidsen - davidsen@sixhub.uucp (uunet!crdgw1!sixhub!davidsen) sysop *IX BBS and Public Access UNIX moderator of comp.binaries.ibm.pc "Getting old is bad, but it beats the hell out of the alternative" -anon
srg@cunixd.cc.columbia.edu (Steven R Gerber) (03/18/90)
Preface: This is not a flame war. Phil, let's move the discussion though. I just got some mail requesting we do. Please e-mail me - I want to keep going. In article <29511@amdcad.AMD.COM> phil@pepsi.AMD.COM (Phil Ngai) writes: >In article <1990Mar15.152746.21025@cunixf.cc.columbia.edu> srg@cunixd.cc.columbia.edu (Steven R Gerber) writes: >|>Be serious. I use both DOS and Unix daily and I have a much better >|>history mechanism on DOS than I do on Unix. >| Did you mean you're using an additional utility (4DOS, stack, etc.)? >| If you mean the regular F3 etc., you gotta be kidding! > >Yes, I have some additional utilities on DOS. I do on Unix too, >things like readnews, emacs, ksh, etc. > >|>I have job control and backgrounding too. You don't HAVE to have Unix >|>to do this. >| True. You can kludge this (Windows, DesqView, etc.). >| But FACE IT: MS-DOS just plain ain't got it. > >I refuse to accept comparisons of DOS without any utilities unless >you also look at Unix without utilities like csh, X, etc. I gotta take this one. The difference here is the OS. Unix has multi-tasking built in - MS-DOS doesn't. This has nothing to do with the shell and/or utilities. BTW Specifically, MS-DOS is NOT re-entrant which makes multi-tasking VERY difficult. > >| I don't want to do back-flips to code. BTW, have you every programmed >| for Windows? It's a real pain-in-the-ass and the SDK is expensive. > >Have you ever programmed for X? Yes. It's a lot easier than Windows. It's portable and it works. Windows and the SDK are almost constantly coming out with new versions which are incompatible with previous. > >Are you saying Unix is not expensive? Certainly it isn't at first. But, after you add all YOUR utilities to MS-DOS the prices go up. Let's take an example (possibly a bad one). ESIX (Everex's unix) is ~$600.00 US including X-Tools. Just MS-Windows with the SDK is about that much. > >-- >Phil Ngai, phil@amd.com {uunet,decwrl,ucbvax}!amdcad!phil >Boycott the census! With the history of abuse census data has, >do you really trust the government? **************************************************************** * Steven R. Gerber - PAL (Programmer At Large) * srg@cunixd.cc.columbia.edu * Tel: 212-794-8721 * UUCP: ...rutgers!columbia!cunixd!srg * FAX: 212-794-8722 ****************************************************************
davidsen@sixhub.UUCP (Wm E. Davidsen Jr) (03/18/90)
In article <1990Mar17.220703.16933@cunixf.cc.columbia.edu> srg@cunixd.cc.columbia.edu (Steven R Gerber) writes: | Preface: This is not a flame war. | Phil, let's move the discussion though. I just got some mail requesting we do. | Please e-mail me - I want to keep going. I have suggested comp.unix.questions, since not everyone gets alt.religion.computers, although that group is just for discussions like this. -- bill davidsen - davidsen@sixhub.uucp (uunet!crdgw1!sixhub!davidsen) sysop *IX BBS and Public Access UNIX moderator of comp.binaries.ibm.pc "Getting old is bad, but it beats the hell out of the alternative" -anon
kjh@pollux.usc.edu (Kenneth J. Hendrickson) (03/18/90)
In article <18459@boulder.Colorado.EDU> wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) writes: >One last thing, rumor has it that Bill Gates is trying to get OS/2 >rewritten in C so that it can be easily transported to new machines. > Dean Wallwey What language did they write OS/2 in? Why wouldn't they write it in C? P.S. I used to have the utmost respect for Microsoft. After learning unix a short time ago, I can now see that MS-DOS is a terrible kludge. Why can't Microsoft just do it right: make the price on Xenix comparable to the price of MS-DOS, and let MS-DOS die the terrible death that it deserves? God knows that Microsoft could make lots of money selling Xenix directly. Just think of all the bucks that could be made selling Xenix to _everybody_, and selling _everybody_ new versions of software they already have that run under Xenix. Ken Hendrickson N8DGN/6 kjh@usc.edu ...!uunet!usc!pollux!kjh
wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) (03/19/90)
In article <1990Mar16.222205.9749@comm.WANG.COM> lws@comm.WANG.COM (Lyle Seaman) writes: >..... >These examples are not nearly as useful as the original proposed example. >So I repeat: Is it capable of keeping an accurate file transfer running >in the background using one serial port while another serial port is also >in use? Easily!!! > >Besides, OS/2 is just warmed-over MS-DOS. I'd rather have TRS-DOS 4.0. > At least from what I have read, OS/2 is completely different. Its a much more "clean" operating system. MS-DOS and even *NIX are both very "kludgy", everything is a "hack job". In OS/2, everything is built right in the system. I know these aren't very quantitative or tangable judgements, but I think they are valid. Dean Wallwey
schaut@cat9.cs.wisc.edu (Rick Schaut) (03/21/90)
In article <6508@skinner.nprdc.arpa> malloy@nprdc.arpa (Sean Malloy) writes: | Unix gets 1/2 because it works well with loose-coupled CPUs but needs | an extension to work with tightly-coupled CPUs. OS/2 gets 1 because it | works well with tightly-coupled CPUs but needs an extension to work | with loosely-coupled CPUs. Is there some kind of bias here you're not | telling us about? Simply a bias towards standards (I _did_ use that word). The LAN Manager is a standardized add-on. So far, AT&T hasn't published _anything_ that will enable UNIX to handle tightly-coupled CPU's. In that sense, the LAN Manager is a part of OS/2 that users can opt not to buy. -- Rick (schaut@garfield.cs.wisc.edu) "I'm a theory geek; we use Turing machines!"--Gary Lewandowski
chris@utgard.uucp (Chris Anderson) (03/22/90)
In article <4492@daffy.cs.wisc.edu> schaut@cat9.cs.wisc.edu (Rick Schaut) writes: >Simply a bias towards standards (I _did_ use that word). The LAN Manager >is a standardized addon. So far, AT&T hasn't published _anything_ that >will enable UNIX to handle tightlycoupled CPU's. In that sense, the LAN >Manager is a part of OS/2 that users can opt not to buy. What about 3b4000's? They have tightly coupled CPU's and run SVR3. This whole bit about tightly coupled CPU's is ridiculous... There are plenty of manufacturers out there who use unix on machines with tightly coupled CPU's. AT&T is *not* the last word in Unix. Nor do they define what Unix is or isn't. That's one of the reasons that OSF was created, to allow other manufacturers to have some say as to what Unix "is". Chris -- | Chris Anderson | | QMA, Inc. email : {csusac,sactoh0}!utgard!chris | |----------------------------------------------------------------------| | My employer never listens to me, so why should he care what I say? |
rommel@lan.informatik.tu-muenchen.dbp.de (Kai-Uwe Rommel) (03/22/90)
Please keep this discussion about OS/2 vs. Unix out of this group. When I look at the articles, many of you talking about how pretty Unix is probably never used (or programmed !) an OS/2 system ! NOBODY wants to forbid you using Unix and NOBODY wants to enforce you to use an OS/2 machine ! Well, besides this flame, here is my opinion: I favour OS/2 MUCH over Unix (I work with Unix daily too, I know what I am talking about !) because its much more modern than Unix. It has VERY fast console I/O, all the things I need under Unix are available too. Over Unix it has Dynamic linking (SCO Unix also has some shard libraries it but nobody uses it) and it has threads (try to run several procedures in ONE module, i.e one source text at the same time on the same data under Unix :-). Compare the size of a non-trivial PM application to a "Hello world!" for X11, then you will know what disk space dynamic linking saves! And, most important for me, OS/2 has a well defined and very nice programming interface ! And what about performance ? - I had a chance of testing both SCO Unix and OS/2 1.1 on the same machine (386, 24 MHz, 8 MB RAM, 80 MB disk). Although X11 runs very well on this machine, it is MUCH slower than the OS/2 PM and even slower than MS-Windows on DOS. The file system was a "AT&T fast file system" (am I right ?) which was even slower than a FAT file system on doing an "rm -r" on a big tree. Also SCO Unix hogs the whole 80 MB disk for system files and programs - where OS/2 for the same functionality (with C compiler, lots of tools, libraries and so on) occupies about 20-25 MB ! Kai Uwe Rommel Munich rommel@lan.informatik.tu-muenchen.dbp.de
lowey@herald.usask.ca (Kevin Lowey) (03/22/90)
From article <1990Mar15.152746.21025@cunixf.cc.columbia.edu>, by srg@cunixd.cc.columbia.edu (Steven R Gerber): >>Be serious. I use both DOS and Unix daily and I have a much better >>history mechanism on DOS than I do on Unix. > Did you mean you're using an additional utility (4DOS, stack, etc.)? > If you mean the regular F3 etc., you gotta be kidding! In OS/2 version 1.2, there is a history mechanism built in which works similar to the popular CED utility for MS-DOS. MUCH nicer than the stupid "history" program in Unix. >>I have job control and backgrounding too. You don't HAVE to have Unix >>to do this. > True. You can kludge this (Windows, DesqView, etc.). > But FACE IT: MS-DOS just plain ain't got it. > I don't want to do back-flips to code. BTW, have you every programmed > for Windows? It's a real pain-in-the-ass and the SDK is expensive. I agree there, but the topic for this message is UNIX vs OS/2, not UNIX vs MS-DOS. OS/2 is a full pre-emptive multitasking operating system, and not a kludge like most of the MS-DOS multitaskers. OS/2 has a number of advantages over Unix for single user machines. First of all, it is designed for low powered machines like my IBM model 60 with a 40M hard disk. To get an equivalent Unix system, you would require not only Unix, but also X-Window. How well would that work on my Model-60? This problem is compounded when you start adding more users on. To get an equivalent GUI interface, my model 60 would have to run unix with x-window locally, and each of the other users would also have to run at LEAST an X-Window terminal. One X-Terminal costs the same as my model-60, with OS/2. So why would I want to run 5 people off my model 60 using unix, when for the same price I could have five independant machines, each dedicated to ONE user. If I want, I can cheaply link these together with a LAN to get file sharing and electronic mail abilities similar to what Unix offers. Heck, if I want to, I can even get Usenet mail feeds on my OS/2 system! - Kevin Lowey
schaut@cat9.cs.wisc.edu (Rick Schaut) (03/22/90)
Note the newsgroups and followup-to lines in the header. I'm directing this discussion to the newsgroup in which it belongs. In article <1990Mar16.222205.9749@comm.WANG.COM> lws@comm.WANG.COM (Lyle Seaman) writes: | wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) writes: | [An example of OS/2 multi-tasking capabilties] | | These examples are not nearly as useful as the original proposed example. | So I repeat: Is it capable of keeping an accurate file transfer running | in the background using one serial port while another serial port is also | in use? 1) To whom are those examples "not nearly as useful"? For the average user, the ability to get information and use it at the same time is more important than the ability to collect information from two different sources at the same time. You've assumed a context that may not be representative of the way that most people use computers. 2) If you want to be able to, say, transfer a file and print another one at the same time, I can do that with Windows on top of good old DOS (I just did it!). Better yet, with OS/2 on a tightly-coupled multiprocessor system, I can do that and still have the _full_ capabilities of another processor available for a third application I might want to run. Neither the AT&T nor the Berkely flavors of UNIX can handle tightly-coupled multi-processors. | Besides, OS/2 is just warmed-over MS-DOS. I'd rather have TRS-DOS 4.0. Why don't you try using a system before flaming it? -- Rick (schaut@garfield.cs.wisc.edu) "I'm a theory geek; we use Turing machines!"--Gary Lewandowski
schaut@cat9.cs.wisc.edu (Rick Schaut) (03/22/90)
Note the newsgroups and followup-to lines in the header. This discussion belongs in c.s.i.p more than it belongs in c.b.i.p.d. In article <1990Mar21.084532.11387@utgard.uucp> chris@utgard.uucp (Chris Anderson) writes: | What about 3b4000's? They have tightly coupled CPU's and run SVR3. | This whole bit about tightly coupled CPU's is ridiculous... There | are plenty of manufacturers out there who use unix on machines with | tightly coupled CPU's. AT&T is *not* the last word in Unix. Nor do | they define what Unix is or isn't. That's one of the reasons that | OSF was created, to allow other manufacturers to have some say as | to what Unix "is". 1) While one can make UNIX work on a tightly-coupled multi-processor system, without multi-threading the system will not take full advantage of having multiple processors. 2) While AT&T, Berkely, Sun, and a host of others decide on a standard way to implement multi-threading in UNIX, I'll use good old standard OS/2. I prefer not to wait for committees. They are notoriously slow. -- Rick (schaut@garfield.cs.wisc.edu) "I'm a theory geek; we use Turing machines!"--Gary Lewandowski
wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) (03/23/90)
In article <1990Mar22.000324.18577@dvinci.usask.ca> lowey@herald.usask.ca writes: >OS/2 has a number of advantages over Unix for single user machines. First >of all, it is designed for low powered machines like my IBM model 60 with >a 40M hard disk. To get an equivalent Unix system, you would require not >only Unix, but also X-Window. How well would that work on my Model-60? > >- Kevin Lowey Exactly! I have a friend that has a 25MHz, cached 386 machine and uses a 32-bit UNIX with X-windows on a 1024 x 768 SUPER VGA. It really is sad to see a $4500 machine have a $450 pc-xt performance!!!!!! Dean Wallwey
neese@adaptex.UUCP (03/23/90)
>>OS/2 has a number of advantages over Unix for single user machines. First >>of all, it is designed for low powered machines like my IBM model 60 with >>a 40M hard disk. To get an equivalent Unix system, you would require not >>only Unix, but also X-Window. How well would that work on my Model-60? >> >Exactly! I have a friend that has a 25MHz, cached 386 machine and >uses a 32-bit UNIX with X-windows on a 1024 x 768 SUPER VGA. It really is >sad to see a $4500 machine have a $450 pc-xt performance!!!!!! I really tried hard to stay out of this, but this was too much to take. I have run both OS/2 and UNIX. To be usable, both require a lot of memory and about the same amount of hard disk. I can find no hardware resources that the both don't need equal amounts of. It is for this reason I will stick with UNIX. And if you get right down to it, you can do a lot more than OS/2 can with XENIX on the same hardware. I am talking about running applications, which OS/2 has very few, but all of the ones that I have used require a lot more memory than a XENIX application that does the same function, maybe not as pretty, but just as functional. And then you start adding stuff on top of both. That is where things get quite a bit different. I can network UNIX systems together with about 4MB's of ram in my server. It runs fine. But OS/2 Lan Manager requires 4MB to boot and you shouldn't run with less than 8MB's of memory. I know memory is cheap, but it is just a comparison. I am not pro OS/2 or pro UNIX. I believe that both have a place in the market. It all depends on your needs. For instance, I prefer UNIX for development purposes, and I haven't quite found out why I need OS/2, but that is just me. The only reason I jumped into this fray, is due to misinformation or misleading information. Roy Neese Adaptec Central Field Applications Engineer UUCP @ {texbell,attctc}!cpe!adaptex!neese merch!adaptex!neese uunet!swbatl!texbell!merch!adaptex!neese