burton@parcvax.UUCP (09/12/87)
I'd like to posit a different view of the value of OS/2, putting aside any issues of delivery date or such. If I understand the arguments correctly, OS/2 offers no benefits over UNIX, particularly UNIX running with a DOS option. I beg to differ, having bet several years of my career on UNIX. Let me say first off that I'm one of those marketing types that so many readers/posters of the net like to make fun of. I am, however, comfortable enough with both DOS and UNIX to understand many of their strengths, even if I don't code in C. There is absolutely no benefit, read NONE, for an operating system itself, except to support applications run by users. And, users will generally pay money for this capability, particularly in the business world where people can justify any kind of office automation on the basis of productivity improvements or increased business. Howeer, most users are funny. They really don't care about the OS, or the chip, or the tracks per inch of the floppy drive. All they want is to get their work done with minimal hassle. And, most people won't/can't learn much about their systems, except to turn it on, load their applications, print results, DOS is popular because it makes very few demands upon the user. This is what most people can cope with. If you don't believe me, watch the "average" user try to cope with the complexities of Lotus or Wordstar. Worse yet, to install wordstar, even if all the necessary instructions are provided. Most people won't/can't do it. And like it or not, folks, these are the people with the money to buy the products you who are developers do. Now, for the average user who is terrified of installing Wordstar or backing up files, or who uses a hard disk without any subdirectories [too much trouble to get them to work right], let's tell that person about UNIX. Any version. Let's tell them about logins, logouts, and supersers. Let's tell them about the hassles in using a floppy with a UNIX system. Compare the simplicity of typing A: to get to the floppy with the RTFM-type syntax of UNIX. Now, let's tell them that the filesystem can easily be corrupted, and that there are stringent requirements for getting it back up. You've lost that person as a paying customer. What's a filesystem? Corruption? (Is that like politics in Boston or Chicago, or is that like AIDS??) Sure, UNIX does more, but the average user today is just getting into LAN's, and doesn't need usenet or mailx or uucp. Or adb or SCCS. I'd love to be the salesman selling against a machine with UNIX, for the average user. (Look at the popularity of the Mackintosh as proof of people's aversion to any kind of command line environment.) And, this isn't just idle ramblings. About five years ago, when PC's were new and 68000's were cheap, companies like Fortune Systems tried to popularize UNIX systems for general office automation use. They even put menus on top of the raw UNIX interface. But it didn't work, even though the economics favored a multi-user Fortune against several standalone PC's. People wanted the simplicity of PC's. They didn't want to have to learn all about a system that was really intended for an environment with professional management. And, in this business, if people don't accept your solution within a year or two, you have been passed by in the market. The marketplace demonstrated that it didn't want the complexities of UNIX-based systems, regardless of the benefits. (The marketplace isn't a few engineers here and there, but the millions and millions of white collar workers who drive cars with automatic transmissions and never built a Heathkit in their lives and use cameras with auto-exposure and automatic flash.) The notion of a UNIX "hypervisor" for multiple virtural DOS as an effective competitor to OS/2 is a vain hope. I personally like using UNIX, and perahps I may even consider it, but I shudder at the thought of UNIX on a machine also used by my wife. or most other average users. So that won't "save" UNIX. --------- I could go on about this, but I hope that readers get the point. Not having seen OS/2, I can't comment about specific points. But, I think that the key issue is that OS/2 will resemble DOS and that it will run today's applications. True, you won't get all the benefits of OS/2 without new applications, but you won't lose too much. The other big issue is that OS/2 is vaporware. I think this is a very myopic perspective. Consider IBM's dilemma. They ahve this evolutionary/ revolutionary operating system that is designed to use the power of the PS/2's (and AT's). The OS is useless without applications. However, IBM doesn't provide any signiciant applications, the OS and the LAN programs excepted. Third parties do. Given today's environment, there is absolutely no way that you can distribute the developer's version of OS/2 and keep it a secret. If you tried that, you would get all sorts of lurid rumors in PC Week about OS/2, AND YOU WOULD CONFUSE YOUR CUSTOMERS. Who would then continue to leave you for DEC and Apple. Much better to tell everyone about it. IBM's real customer base in the Fortune 1000 is quite sophisticated about the development process. Small busienss doesn't care that much, and individual users just aren't that much of a factor. So, sports fans, who's on first?? [What's on second, and I-don't-give-a-damn is on third -)]. -- Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton Xerox Corp. preferred path: burton.osbunorth@xerox.COM 408 737 4635 ... usual disclaimers apply ...
farren@hoptoad.UUCP (09/13/87)
In article <494@parcvax.Xerox.COM> burton@parcvax.Xerox.COM writes: > >Now, for the average user who is terrified of installing Wordstar or backing >up files, or who uses a hard disk without any subdirectories [too much trouble >to get them to work right], let's tell that person about UNIX. Any version. > >Let's tell them about logins, logouts, and supersers. Let's tell them about >the hassles in using a floppy with a UNIX system. Compare the simplicity of >typing A: to get to the floppy with the RTFM-type syntax of UNIX. > >Now, let's tell them that the filesystem can easily be corrupted, and that >there are stringent requirements for getting it back up. You've lost that >person as a paying customer. What's a filesystem? Corruption? (Is that >like politics in Boston or Chicago, or is that like AIDS??) > No, you don't tell them any of those things. You set up a system, as a shell, which is as easy to use as MS-DOS (and probably a lot easier), but which also allows access to deeper, more cryptic levels of Unix if they want them. You provide, within that system, the tools which are necessary to let them do their work, and you make sure that the tools are easy-to-use. You provide, within your Unix implementation, as much protection as possible against the types of file corruption that you cite (it isn't easy, but it's possible). Then, you let them go. Unix, in its history, has been viewed as an OS for the computer sophisticate. In part, this is true, in that to use all of its capabilities requires a fairly large amount of knowledge of its structure. The "normal" user, however, doesn't need to do this type of access at all, and Unix is very nicely set up to allow differing types of access for differing types of needs. All it takes is a little work, and, unfortunately, few people seem to be very interested in doing this work. "Well, *I* understand it, so why can't they?" seems to be a prevalent attitude, and it is WRONG. For an example of how Unix can make things easier, rather than more complex, I recommend looking at some of Larry Wall's shell Configure scripts, where he has automated the process of installing some of his very complex pieces of software to the point where it is rather difficult to make it NOT work right. Within his scripts, he determines the type of Unix you are running, the capabilities of your system, where some standard utilities live, and more, all without requiring the person running the script to understand a thing about any of it. A similar capability can, and probably should, be applied at the shell level; sadly, noone has done it yet. In my opinion, the whole concept of "Unix is a complex, difficult, and cryptic OS, which ordinary mortals cannot possibly understand" is a crock. If ordinary mortals can't understand it, then it is the fault of us extra-ordinary mortals. We can, and should, make it possible to use Unix as easily as using MS-DOS, or even as easily as using the Mac. -- ---------------- "... if the church put in half the time on covetousness Mike Farren that it does on lust, this would be a better world ..." hoptoad!farren Garrison Keillor, "Lake Wobegon Days"
brad@looking.UUCP (09/13/87)
My compliments on an insightful and mostly correct commentary. But there is more, and this gives Unix a fighting chance: Unix's command structure can be altered to become less frightening. Menus etc. are not the answer, as Fortune found out. The main reason Unix hasn't changed to a more consistent command structure is resistance from the Unix hacks. Oddly, it is the same unix hacks who decry the resistance to change of conservative DOS users! The biggest problem is the difficulty in administering and maintaining a Unix system. It is still orders of magnitude more difficult than the "turn it on and run a program" simplicity of DOS. Many of these problems are not specific to Unix, but are caused by having dynamic, multi-tasking, multi-user networked environments. OS/2 will fall plague to some of these. The big advantage Unix has is that Unix for the 80386 is available now. OS/2 for the 386 is predicted by Microsoft to be over a year away. If the delivery is anything like that for MS Windows, well.... In the interim people writing 386 software will have only one OS to run it under. Thus all new 386 applications will be developed for Unix (and other, non-OS/2 systems that come on-line) Thus Unix will be the system that runs the old software and the new, and this will help it overcome the other perceived troubles. The alternative to this is that the 386 suffers the fate of the 286 -- simply being used as a fast version of its predecessor. I hope we don't see this. -- Brad Templeton, Looking Glass Software Ltd. - Waterloo, Ontario 519/884-7473
burton@parcvax.Xerox.COM (Philip M. Burton) (09/13/87)
In article <2944@hoptoad.uucp> farren@hoptoad.UUCP (Mike Farren) writes: >In article <494@parcvax.Xerox.COM> burton@parcvax.Xerox.COM writes: >> >>Now, for the average user who is terrified of installing Wordstar or backing >> >No, you don't tell them any of those things. You set up a system, as a >shell, which is as easy to use as MS-DOS (and probably a lot easier), but >which also allows access to deeper, more cryptic levels of Unix if they >want them. You provide, within that system, the tools which are necessary >to let them do their work, and you make sure that the tools are easy-to-use. >You provide, within your Unix implementation, as much protection as possible >against the types of file corruption that you cite (it isn't easy, but it's >possible). Then, you let them go. > Sorry to throw cold water, but that approach has already been tried by Fortune Systems, AT&T and probably others. Unfortunately, it's a lot of work to get there, and you still have the problems of mounting a floppy, e.r. disk pack, and still risk filesystem corruption if you don't shut the machine down properly. etc. etc. Another fundamental problem was highlighted by Bill Gates at a recent meeting of the Silicon Valley Computer Society recently. UNIX has a model of dumb terminals communicating via serial lines to a host. DOS and OS/2 assume a much tighter linkage. How many serial line-connected UNIX machines actually support bit-mapped graphics? Several years ago Convergent Technologies had such a beast, and it was connected via an RS-422 (I think) at 307 Kbps!!! Don't misunderstand. I think the world would be a much better place if we could provide "DOS under UNIX", if only because then we wold have a solution that was not restricted to Intel 286/386 systems. (Imagine running your DOS applications on a big, big VAX or on a Cray.) But it ain't gonna happen. (wasted space for my stupid mailer) -- Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton Xerox Corp. preferred path: burton.osbunorth@xerox.COM 408 737 4635 ... usual disclaimers apply ...
richard@gryphon.UUCP (09/13/87)
In article <494@parcvax.Xerox.COM> burton@parcvax.Xerox.COM (Philip M. Burton) writes: >I'd like to posit a different view of the value of OS/2, putting aside any >issues of delivery date or such. > >If I understand the arguments correctly, OS/2 offers no benefits over UNIX, >particularly UNIX running with a DOS option. I beg to differ, having bet >several years of my career on UNIX. Did the gamble pay off :-) >Let me say first off that I'm one of those marketing types that so many >readers/posters of the net like to make fun of. I am, however, comfortable >enough with both DOS and UNIX to understand many of their strengths, even >if I don't code in C. > >There is absolutely no benefit, read NONE, for an operating system itself, >except to support applications run by users. And, users will generally pay >money for this capability, particularly in the business world where people >can justify any kind of office automation on the basis of productivity >improvements or increased business. From a marketing point of view this is true, you sell solutions. But when the user is comfortable using his solution (WS or 123, whatever) and actually discovers how to use the operating system commands, they quickly run up against limitations in RT/11 err, CP/...ERR MS-DOS. >Howeer, most users are funny. They really don't care about the OS, or the chip, >or the tracks per inch of the floppy drive. All they want is to get their >work done with minimal hassle. And, most people won't/can't learn much about >their systems, except to turn it on, load their applications, print results, > >DOS is popular because it makes very few demands upon the user. This is what >most people can cope with. If you don't believe me, watch the "average" user >try to cope with the complexities of Lotus or Wordstar. Worse yet, to install >wordstar, even if all the necessary instructions are provided. Most people >won't/can't do it. And like it or not, folks, these are the people with the >money to buy the products you who are developers do. I dunno. Most folks I've seen have no problem with WS or 123, its the Zenographics Mirage and Future-Net type applications that seem to be a pain. >Now, for the average user who is terrified of installing Wordstar or backing >up files, or who uses a hard disk without any subdirectories [too much trouble >to get them to work right], let's tell that person about UNIX. Any version. >Let's tell them about logins, logouts, and supersers. Let's tell them about Installing wordstar ? 'copy ws*.* c:' Screw the installation instructions. :-) >the hassles in using a floppy with a UNIX system. Compare the simplicity of >typing A: to get to the floppy with the RTFM-type syntax of UNIX. > >Now, let's tell them that the filesystem can easily be corrupted, and that >there are stringent requirements for getting it back up. You've lost that >person as a paying customer. What's a filesystem? Corruption? (Is that >like politics in Boston or Chicago, or is that like AIDS??) I've corrupeted disks from here to Cucamonga, and have my share of PC disks that are un-recoverable. Sorry, the only corrupted filesystem I've found I can recover 100% is the Amiga's (No flames here about the inapproprtaeness of the amiga to run WordStar, I'm only talking about the integrity of the filesystem when any given track is trashed) >Sure, UNIX does more, but the average user today is just getting into LAN's, >and doesn't need usenet or mailx or uucp. Or adb or SCCS. I'd love to be >the salesman selling against a machine with UNIX, for the average user. >(Look at the popularity of the Mackintosh as proof of people's aversion to >any kind of command line environment.) Re: the average user today getting into LAN's. Well, golly, I don't know much about all this, but for the year I used a 3Com network (when it worked) I felt like rewriting uucp would be easier than dealing with limited tools 3Com gave us. >And, this isn't just idle ramblings. Mine are. >About five years ago, when PC's were new >and 68000's were cheap, companies like Fortune Systems tried to popularize >UNIX systems for general office automation use. They even put menus on top >of the raw UNIX interface. I wonder what would have happened if you took one of those fortune systems, painted it blue or beige and put a big fat IBM label on it ? >But it didn't work, even though the economics favored a multi-user Fortune >against several standalone PC's. People wanted the simplicity of PC's. They >didn't want to have to learn all about a system that was really intended for >an environment with professional management. Peoeple didn't want the simplicity of PC's, they could have had that with RT/11 or CP/M, what they wanted was THEIR OWN (IBM) COMPUTER to sit on THEIR DESK. >And, in this business, if people don't accept your solution within a year or >two, you have been passed by in the market. The marketplace demonstrated >that it didn't want the complexities of UNIX-based systems, regardless of the >benefits. (The marketplace isn't a few engineers here and there, but the >millions and millions of white collar workers who drive cars with automatic >transmissions and never built a Heathkit in their lives and use cameras with >auto-exposure and automatic flash.) Was it really the solution that sold them, or the marketing and company image ? >The notion of a UNIX "hypervisor" for multiple virtural DOS as an effective >competitor to OS/2 is a vain hope. I personally like using UNIX, and perahps >I may even consider it, but I shudder at the thought of UNIX on a machine also >used by my wife. or most other average users. So that won't "save" UNIX. ^^^^ Hope Cheryl doesn't hear this obvious denegration of women :-) >--------- > >So, sports fans, who's on first?? [What's on second, and I-don't-give-a-damn >is on third -)]. > Who's on first, Guess Who's on second, Yes is on third. >-- >Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton -- Richard J. Sexton INTERNET: richard@gryphon.CTS.COM UUCP: {hplabs!hp-sdd, sdcsvax, ihnp4, nosc}!crash!gryphon!richard "It's too dark to put the key in my ignition..."
burton@parcvax.Xerox.COM (Philip M. Burton) (09/13/87)
In article <961@looking.UUCP> brad@looking.UUCP (Brad Templeton) writes: >My compliments on an insightful and mostly correct commentary. But there >is more, and this gives Unix a fighting chance: > Thanks. Unfortunately, I came to that conclusion reluctantly. I wish I could decided otherwise. Don't forget that IBM has had several years of selling UNIX on mainframes, on XT's and AT's and with RT's, and you can be sure that they have had people studying it in their research labs. However, I feel that if I were the product planner for the OS/2, I probably would have made the decision NOT to use UNIX. > Unix's command structure can be altered to become less frightening. > Menus etc. are not the answer, as Fortune found out. The main reason > Unix hasn't changed to a more consistent command structure is resistance > from the Unix hacks. > I agree. One of the less-afflicted hack types at Fortune, and believe me, we had some of the best developers around, said that, hey no big deal, just use mv to rename a file. Right!! Just tell that to Joe user. However, menus are part of the answer. For many people, any kind of command line is NOT the answer. That's a good part of the success of the Mac. > > The biggest problem is the difficulty in administering and maintaining a > Unix system. It is still orders of magnitude more difficult than the > "turn it on and run a program" simplicity of DOS. Many of these problems > are not specific to Unix, but are caused by having dynamic, multi-tasking, > multi-user networked environments. OS/2 will fall plague to some of these. > I have to agree and disagree. Any multi-user system would have many of these problems. Could you imagine, in your wildest dreams, putting IBM's MVS on a PC. I'm sure that a 386 would have the horsepower to do it, but who would want JCL and its horrible cousins. At the same time, until you make fundamental, and I do mean that, changes to UNIX, for example the filesystem fragility, lack of true random access, you don't have a PERSONAL operating system. Now, if you did make all those changes, would it still be UNIX??? > The big advantage Unix has is that Unix for the 80386 is available now. > OS/2 for the 386 is predicted by Microsoft to be over a year away. If > the delivery is anything like that for MS Windows, well.... > In 1981, the only software for IBM PC's was CP/M based, and people actually went out and bought CP/M co-processor boards for the PC's. The early public domain club software disks were all CP/M reworks. But by 1983, interest in CP/M was gone from the mass market. In 1984, I worked briefly for a company that needed a CP/M co-processor card, and by then, the CP/M co-processor card was moribund. > In the interim people writing 386 software will have only one OS to > run it under. Thus all new 386 applications will be developed for Unix > (and other, non-OS/2 systems that come on-line) See my comments above about CP/M. > > Thus Unix will be the system that runs the old software and the new, > and this will help it overcome the other perceived troubles. > >The alternative to this is that the 386 suffers the fate of the 286 -- simply >being used as a fast version of its predecessor. I hope we don't see this. > >-- >Brad Templeton, Looking Glass Software Ltd. - Waterloo, Ontario 519/884-7473 No, that conclusion about UNIX doesn't follow. UNIX will not run software written for OS/2. And even if it were true, my "marketing gut" tells me that the advantages of UNIX over OS/2, even if true, will not be enough to sway most people in the great mass market from going to OS/2. I'm sure that IBM will "fix" the problem with a native 386 version of OS/2. However, I recently heard that the 386 isn't the real long-term solution because it can't run multiple virtual versions of itself. And, the REAL solution was the followon chip, which would allow multiple virtual OS/2's. IBM will certainly bring out a 386 native version of OS/2, when it's time to obsolete the AT's and all the 286 machines on the market, or at least differentiate their model 80 (and its unannounced cousins) from other systems. I'm going to hazard a guess and estimate a mid-89 announcement for OS/2 for the 386. Sorry, guys. I like UNIX. I really do. What got me started on all this was the prospect that MicroPort has a UNIX that I could actually afford to buy just to hack around with. Remember I'm not a developer, so it wold be mainly a toy to play with. Then I started thinking of the incredible hassles that I would have to put up with to get uucp news up, or whatever. Just not worth it. And if I don't want to do it, and I'm faily technical for a non-developer (at least the guys at work tell me that), then it's a hopeless situation for the person who buys a PC just to do spreadsheets of maintain a database. -- Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton Xerox Corp. preferred path: burton.osbunorth@xerox.COM 408 737 4635 ... usual disclaimers apply ...
rcw@qetzal.UUCP (09/13/87)
In article <494@parcvax.Xerox.COM>, burton@parcvax.Xerox.COM (Philip M. Burton) writes: > [much stuff about how DOS is so much simpler to use deleted] Customer Question: Can two or more people use my DOS application simultaneously? Me: Unfortunately not. Let me tell you about Unix(TM)... Customer Question: What about going with OS/2 Version 1.0? Me: Never go with Version 1.0 of anything... ----- In general, I have to agree with burton@parcvax.Xerox.COM's observations. When writing a new application, it may first be delivered on DOS. When the user becomes familiar with that application, it is then much easier to move it to UNIX(TM) and train the customer how to access his application. If done correctly, and you allow training time in the contract, the user will be extremely happy with the multiuser implementation of his application. In some applications, DOS provides a viable upgrade path. In terms of training neophyte users, we have been doing this successfully for years in the oil and gas business. The secret is to have lots of patience and personal contact with the user. Are we comparing apples (DOS) to oranges (UNIX(TM)) here? I think so. For some applications, DOS would never be considered. For example, an application which requires two or more people accessing a database at the same time. On the other hand, I would not consider UNIX(TM) for a single person spreadsheet application. A user's willingness to learn more about her/his system is also a big factor. If an application is very critical, I find they are more willing to learn whatever is necessary to use the application. The difficulty of using a UNIX(TM) application is more of a reflection on the designer than on the user. It is very possible to "hide" most of the Unix crap from the user. It is not trivial to do. Because of closeness and familiarity with the OS, you can forget to "hide" it properly. Let a user review it for you, and listen to the suggestions. OS/2 will be faced with many of the issues Unix faces. From the rumors I hear on the net, it appears that OS/2 will be very unix-like. For some reason, the belief persists that a multiuser operating system is just an enhanced DOS. It is not. It is much more complex. If you can think of a way to implement a multiuser operating system that is as simple to use as DOS, do it. You will be rich. -- Robert C. White, Jr *---- ----* MENTOR SOFTWARE, INC. 1-303-252-9090 ihnp4!upba!qetzal!rcw | |_| | Specializing in Unix applications, 11534 Steele St. | _ o _ | system software and administrativa. Thornton, CO 80233 | | | | | | "We will make it work for you."
rcw@qetzal.UUCP (09/13/87)
In article <961@looking.UUCP>, brad@looking.UUCP (Brad Templeton) writes: | The biggest problem is the difficulty in administering and maintaining a | Unix system. It is still orders of magnitude more difficult than the | "turn it on and run a program" simplicity of DOS. Many of these problems Much as I love to flame the IBM-RT, IBM has done a very good job of hiding the boot sequence. You literally turn the box on, and in a few minutes it says "login:". You don't have to sit through fsck's POSSIBLE FILE SIZE error messages, or get panicky phone calls from users. The diagnostic diskette idea is slick too. Avoids the [it's a hardware problem - no it's a software problem] conflict for the most part. I also like how the RT figures out what devices are connected to it. With my AT clone, I have had to battle every component ever purchased for it. With the RT, you just stick it in, and it figures it out for you. Configuration can be displayed using the "devices" command. The best of the best is the digitizing tablet. No switches! Back in the bad old days, it always took the better part of half a day to get any serial device to work, and always had to break out the black box. Not so with the RT. Just plug it in, and away it goes. | The alternative to this is that the 386 suffers the fate of the 286 -- simply | being used as a fast version of its predecessor. I hope we don't see this. Amen! -- Robert C. White, Jr *---- ----* MENTOR SOFTWARE, INC. 1-303-252-9090 ihnp4!upba!qetzal!rcw | |_| | Specializing in Unix applications, 11534 Steele St. | _ o _ | system software and administrativa. Thornton, CO 80233 | | | | | | "We will make it work for you."
burton@parcvax.Xerox.COM (Philip M. Burton) (09/14/87)
In article <1522@gryphon.CTS.COM> richard@gryphon.CTS.COM (Richard Sexton) writes: >In article <494@parcvax.Xerox.COM> burton@parcvax.Xerox.COM (Philip M. Burton) writes: >>I'd like to posit a different view of the value of OS/2, putting aside any >>several years of my career on UNIX. > >Did the gamble pay off :-) > That all depends on how your view a gamble. Certainly, in of itself, it was a lousy situation. Stock option always "underwater," the good people at the company (many are still net readers and you know who you are) got the old shafteroo, but lots of politics and get-rich-quick. But, as an education, it sure beats most of my formal education, in terms of how poor management can sink a good product, which it was for the time. And, I'll never again believe the promises of how great this is going to be, no matter how sincere the people. > >From a marketing point of view this is true, you sell solutions. But >when the user is comfortable using his solution (WS or 123, whatever) >and actually discovers how to use the operating system commands, I know CP/M and DOS, not RT/11, so I'll speak only of those. Their limitations are frustrating, but the limitations generally don't keep people from using the system as a "plug and play" box. Imagine if you had to get a ham radio licence just to play your stereo or a TV technician's training just to use your VCR. That is what UNIX demands. Too much complexity just to get started. >Installing wordstar ? 'copy ws*.* c:' Screw the installation instructions. :-) Then you obviously never installed Wordstar. It has a huge number of options, many of which require a good understanding of your printer type (you may need to enter printer control codes) and your port type. Further, there is a whole "secret" menu that is really essential in setting up defaults like line length. > >>the hassles in using a floppy with a UNIX system. Compare the simplicity of >>typing A: to get to the floppy with the RTFM-type syntax of UNIX. >> > >I've corrupeted disks from here to Cucamonga, and have my share of PC Are you confusing bad media with actual filesystem corruption? Almost all of my DOS filesystem problems are media related. Same was true in my CP/M days, except when I've done some hacking with Norton Utilities or the CP/M DU program. >Re: the average user today getting into LAN's. Well, golly, I don't >know much about all this, but for the year I used a 3Com network >(when it worked) I felt like rewriting uucp would be easier than >dealing with limited tools 3Com gave us. And guess what. Most users still feel and 3COM is the high end of LANs, and they don't need something as "powerful". >>About five years ago, when PC's were new >>and 68000's were cheap, companies like Fortune Systems tried to popularize >>UNIX systems for general office automation use. They even put menus on top >>of the raw UNIX interface. > >I wonder what would have happened if you took one of those fortune systems, >painted it blue or beige and put a big fat IBM label on it ? > The Fortune 32:16 would have sold one hell of a lot more systems if it had been called the IBM PC VAT or some such. No questions about that. But then again, speaking from personal experience, IBM treats its customers better than Fortune did. [No flames, and I can't comment back, unless I want the founders of Fortune to sue me.] >Peoeple didn't want the simplicity of PC's, they could have had that with >RT/11 or CP/M, what they wanted was THEIR OWN (IBM) COMPUTER >to sit on THEIR DESK. Wrong!!! They wanted a box that could run Lotus 1-2-3. Just as an earlier generation of buyers didn't want an Apple II, but a box that ran Visicalc. Another thing to remember is that most of the old CP/M companies were very poorly managed, although Adam Osborne's debacle was more visible. The dealers hated most of their CP/M company vendors, and they welcomed IBM because of it's obvious name recognition. But it was the dealers who helped make IBM, because in the retail computer business, like any retail business, "shelf space" will make or break any product. > >>transmissions and never built a Heathkit in their lives and use cameras with >>auto-exposure and automatic flash.) > >Was it really the solution that sold them, or the marketing and company >image ? Well, it is obvious that marketing and image has had lots to do with the success of the IBM PC. But isn't that true of cars, cameras, perfumes, food fads, politics, styles in dress, or most other human activities. There no philosopher king who makes the decisions. >>I may even consider it, but I shudder at the thought of UNIX on a machine also >>used by my wife. or most other average users. So that won't "save" UNIX. > ^^^^ >Hope Cheryl doesn't hear this obvious denegration of women :-) I don't know who Cheryl is, but my wife Ilene would not be. She tells everyone she can about how this stuff isn't made for her, but techie-types like me. My wife is no dummy, but she simply doesn't care about this stuff. She simply wants to use the damn thing to run her business. >Richard J. Sexton [wasted space for the stupid mailer.] -- Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton Xerox Corp. preferred path: burton.osbunorth@xerox.COM 408 737 4635 ... usual disclaimers apply ...
tankus@hsi.UUCP (Ed Tankus) (09/14/87)
Phillip Barton made some interesting comments. His point of view is from the
marketing end rather than from the developer/hacker/guru perspective. Each
has a point to be made and each has done it and done it well.
Now the bottom line is this: OS/2, U*IX, or ABCD operating system will NOT
take over the world. Over time, each will have its own market, and hopefully,
enough overlap, that we will start moving towards a more homogenous environment.
This is not possible now and maybe not in 5 years. But let's face it, we will
make our own choices based on what is easier for us as individuals, as
business people, educators, and corporations. Time will tell what operating
system will be prevalent and in which market. Afterall, the marketplace
determines the size and shape and quantity of what sells. The marketplace
determined that the Intel chips were what was needed rather than the more
sophisticated 68000 chips.
The arguing can go on, but lets not continue. I think most will agree that
we have aired enough of our own views to last a few months until the next
discussion on this same subject starts anew.
Thanks to all those who did have something to say. It was very interesting.
Now let's move on.
Cheers!
-- Ed.
Net : {uunet,ihnp4,noao,hao,yale}!hsi!tankus
Snail: Health Systems Int'l, 100 Broadway, New Haven, CT 06511
Bell : (203) 562-2101acm@bu-cs.BU.EDU (ACM) (09/14/87)
In article <494@parcvax.Xerox.COM> burton@parcvax.Xerox.COM (Philip M. Burton) writes: >Howeer, most users are funny. They really don't care about the OS, or the chip, >or the tracks per inch of the floppy drive. All they want is to get their >work done with minimal hassle. And, most people won't/can't learn much about >their systems, except to turn it on, load their applications, print results, > >DOS is popular because it makes very few demands upon the user. [...] > >Now, for the average user who is terrified of installing Wordstar or backing >up files, or who uses a hard disk without any subdirectories [too much trouble >to get them to work right], let's tell that person about UNIX. Any version. > >[...] Compare the simplicity of >typing A: to get to the floppy with the RTFM-type syntax of UNIX. > >Now, let's tell them that the filesystem can easily be corrupted, and that >there are stringent requirements for getting it back up. You've lost that >person as a paying customer. What's a filesystem? Corruption? (Is that >like politics in Boston or Chicago, or is that like AIDS??) He has a fantastic point. Consider my example: I work for a small accounting firm in NH. We recently installed a multiuser PC environment (basically a LAN but really a clustered CPU system). I got the tedious task of explaining to a secretary how to work the system. I ended up making batch files for EVERYTHING. If she wants to do a backup she has to bring the system down (multiuser environments make backups more tricky). I explain how to do this without getting users mad. Now, getting the backup started was more complicated since you had to bring the system up in single-user mode. The best way was to make a batch file and ask her plain-English questions: Do you want to do a backup? [y/n] If she does, the system doesn't go into multiuser mode and instead runs the backup program. Our secretary who does most of the DP work understands; the others can't even handle something this simple. Unfortunately the younger generation that grew up with micros are getting exposed to very single-user systems (eg Commodore-64, Apple //, IBM PC) so they, even though more saavy than their parents, still will have problems with UNIX. I know -- I also work as a terminal assistant in the Comp. Sci. terminal room. I see users every day who say "well my IBM PC doesn't do it that way!" It's too bad. I see the power of UNIX. But I also see that not everyone wants to learn what the machine can do. So you end up with a popular, although braindamaged, OS. What's a poor programmer to do? %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Jim Frost * The Madd Hacker | UUCP: ..!harvard!bu-cs!bucsb!madd H H | ARPA: madd@bucsb.bu.edu H-C-C-OH <- heehee +---------+---------------------------------- H H | "We are strangers in a world we never made"
tim@cit-vax.Caltech.Edu (Timothy L. Kay) (09/14/87)
Philip writes: >Don't forget that IBM has had several years of selling UNIX on mainframes, Really? I have been consulting for IBM Research for several years. I have been asking them to get me a version of Unix for their mainframes the whole time. It doesn't even have to be a product, but simply something I can use in house. Zip. They can't deliver. As long as I have the floor, I might as well add my two cents regarding the Unix vs OS/2 discussion. I have been creating custom database applications for businesses for many years. I have done this using MSDOS because of its availability and inexpense. I am looking forward to the prospect of 80386 and Unix. Multiuser solves the multiple access problem a whole lot better and less expensively than any of the networking that the market has created. I will be setting up the whole show, so installation isn't a problem. I'll be able to do remote maintenance. The only concern is the delicate file system. Stallman says that GNU (a more or less free Unix compatible), will have a fail safe file system. This is possible as long as you are willing to have extra disk space and pay a little in performance. So, maybe somebody (are you listening Microport???) could write a robust filesystem for Unix now. Or, convince GNU to release the filesystem part of the kernel early. Of course, you don't have to worry about them shutting down the machine correctly--I tell them never turn it off. When Unix is running, I might just disable the power switch. Tim
matt@inuxf.UUCP (Matt Verner) (09/15/87)
In article <961@looking.UUCP>, brad@looking.UUCP (Brad Templeton) writes: | The biggest problem is the difficulty in administering and maintaining a | Unix system. It is still orders of magnitude more difficult than the | "turn it on and run a program" simplicity of DOS. Many of these problems It seems to me that when you use the argument that Unix is big and complex (and hard to use/admin) and DOS is simple, you are laying the base of an argument against OS2! OS2 is likely to require 400K of memory and be just as hard, if not harder, to configure for the huge number of devices that are available for the PC/AT. Unix is no picnic for the niave user and I truly can't see OS2 being any better if it has half the power of Unix. Matt Verner UUCP: ...ihnp4!inuxc!matt AT&T Graphics Software Labs AT&T: (317) 844-4364 Indianapolis, IN "The whole point of this sentence is to clearly explain the point this sentence is making."
dboyes@uoregon.UUCP (David Boyes) (09/15/87)
You might want to check out IX/370, a Unix port for the System/370 architecture that runs very nicely as both a standalone system or a dial-in virtual machine under VM/SP. Believe it or not, IBM did a fairly good job -- it can even use 3270 style terminals in a reasonable fashion. Admittedly, it is slow, but translating a lot of this stuff second-level makes anything slow. Disclaimer: I've installed IX/370 once. It's a bitch to install, but now that it's up, it seems to be a resonable system. It's as good as HP-UX ever was.... -- David Boyes ARPA: 556%OREGON1.BITNET@WISCVM.WISC.EDU Systems Division BITNET: 556@OREGON1 University of Oregon Computing Center UUCP: dboyes@uoregon.UUCP
brad@looking.UUCP (09/15/87)
Phillip Burton suggests that the quick death of CP/M, in spite of it's position as the OS for most applications software in 1981, fortells a similar fate for Unix. This was in response to my comment that because Unix would be the only major 386 OS for the next 2 years, true 386 applications would move there, giving it life. I think the situations here are quite a bit different for several reasons: 1) The PC software that dominated was quite a bit better than the CP/M software. CP/M was for the 8080 (although many ran it on Z-80 systems) while PC software had access to the 8086 and even the 8087. In addition, 8080 software could be translated up. The Unix vs OS/2 competition will be fought on the same processors. No inherent reason one bit of software should be superior to another. There will be marketing reasons. 2) The Unix software base is considerably larger and more widely used than the CP/M software base was. It contains a lot of powerful software, once written for large machines, that can now be run on micros. 3) Unix is supported by AT&T, until recently a company larger than IBM. Of course, they don't give it the support their size can truly offer, but in the crunch... Compare this with Digital Research, who were unable to support CP/M enough to get IBM to bless it the way they did MS-DOS. 4) Unix is also supported, in the Xenix variant, by Microsoft, although Microsoft may face pressure from IBM to favour OS/2 over Xenix. 5) Software today is written more portably. Back then, putting your application on DOS and CP/M would have been double the work. Putting it on Unix and OS/2 isn't nearly so bad. This helps OS/2 as well, as old Unix applications can be moved over, but what it really means is that the Unix base doesn't have to be nearly as large before developers can profit by supporting it. 6) (linked of 5) Many companies actually do their DOS development on Unix already. We do, and then move it down to DOS. The Microsoft C compiler under Xenix makes this easy. The 286 protected mode environment is much better for testing software since all bad pointers trap. This means that software done this way can be sold with zero development effort. That's not to say zero marketing, documenting and packaging effort, but it does make it easier and helps break the threshold. Again, the big factor is the 386 and the two year gap while Xenix 386 runs and OS/3(?) is vapour. Corporate micro users are buying 386s because they need power, and because they want the fanciest thing around on *their* desk. For whatever reason, when it comes time for them to run a 386-needing application, they will only have Unix and small players to choose from. Those who can, may wait. But those who can't wait will run Unix. And then Unix's weaknesses will become its strengths. The more they run Unix, the more they will need to STAY with Unix. The harder it is to maintain, the more $$$$ they will invest in maintaining it and the more time they will spend learning to use and maintain it. By the time OS/3 ships, they will not want to change. Thus Unix *can* become the dominant OS, but it needs a) Lots of people who need or want or think they want 386 power b) 386 applications NOW. Part b should be easy, because the 386 is a fine machine to compile C for. All of the old programs for Vaxen and 68000s should come in fine. As noted above, many programs are already developed with Unix. ------------ The Unix command structure and administration structure does need to change. As I noted before, the Unix hacks are the block to this. Right now Unix sells mainly to people already sold on Unix, or people who buy it as part of an OEM package. It loses others because of the reputation Unix has. If you change the commands etc. you lose the old folks and wait a long time to get the new folks. Not profitable. Unix can be changed heavily and still be Unix. It's "Unix" in many respects if it can still run software written for Unix. It doesn't have to look like that on the outside. Unix is a trade mark of Bell Labs. -- Brad Templeton, Looking Glass Software Ltd. - Waterloo, Ontario 519/884-7473
agnew@trwrc.UUCP (R.A. Agnew) (09/15/87)
>benefits. (The marketplace isn't a few engineers here and there, but the >millions and millions of white collar workers who drive cars with automatic >transmissions and never built a Heathkit in their lives and use cameras with >auto-exposure and automatic flash.) > I couldn't agree more! Would all of you millions of white collar workers go off and play with your 123 and OS-2 and leave Unix where it belongs. After all, it was written by engineers, for engineers. (Note the bessel and error functions in the C run time library and the plot and graph routines ). Meanwhile, I'll design your next user friendly PC for you. I'll make sure not to make it too hard to use. Standard disclaimer applies.
burton@parcvax.Xerox.COM (Philip M. Burton) (09/15/87)
In article <680@hsi.UUCP> tankus@hsi.UUCP (Ed Tankus) writes: > >Now the bottom line is this: OS/2, U*IX, or ABCD operating system will NOT >take over the world. Over time, each will have its own market, and hopefully, >enough overlap, that we will start moving towards a more homogenous environment. >This is not possible now and maybe not in 5 years. But let's face it, we will Brave noises from someone who probably doesn't remember the plethora of OS's that existed for micros *for the business world* prior to 1981. There was CP/M, of course, but there was also the P-system. INcidentally, IBM offerred these two in addition to DOS by IBM when the PC first came out. There was also OS-9, the Apple DOS, one or more for various TRS-80 models, and probably some others forgotten by history. >make our own choices based on what is easier for us as individuals, as Only the truly independently wealthy can say that. Others must consider spouses, children, empllyees and employers, and mortgage-holders. >business people, educators, and corporations. Time will tell what operating >system will be prevalent and in which market. Afterall, the marketplace >determines the size and shape and quantity of what sells. The marketplace By now, you shouldhave some notion that various market forces will favor OS/2 over UNIX. (Just ask Lotus or Ashton-Tate what their product plans are. -) -). ) Sure, UNIX will continue to have a "niche" role. And I still like it. But I'm not goin to bet the ranch on it. -- Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton Xerox Corp. preferred path: burton.osbunorth@xerox.COM 408 737 4635 ... usual disclaimers apply ...
dave@westmark.UUCP (Dave Levenson) (09/16/87)
In article <198@qetzal.UUCP>, rcw@qetzal.UUCP (sysop) writes: ... > For some applications, DOS would never be considered. For example, > an application which requires two or more people accessing a database > at the same time. ... We just completed the installation of a multi-user database system under MS-DOS. What we have is multiple MS-DOS systems connected by a lan to a MS-DOS database server. Why did we do it this way? Our customer was already familiar with MS-DOS. The distributed architecture is robust in the face of a system failure. (The database is kept on one of the client workstations, which can take over the server role when necessary.) Yes, it could have been done under UNIX, with everybody sharing the same system, but at a higher price-per-user, for the same performance. -- Dave Levenson Westmark, Inc. A node for news. Warren, NJ USA {rutgers | clyde | mtune | ihnp4}!westmark!dave
dave@westmark.UUCP (Dave Levenson) (09/16/87)
In article <498@parcvax.Xerox.COM>, burton@parcvax.Xerox.COM (Philip M. Burton) writes: > Then I started thinking of the incredible hassles that I would have to put > up with to get uucp news up, or whatever. Just not worth it. > I run UNIX & MS-DOS simultaneously on an AT&T 6300PLUS '286 system. It came with HoneyDanBer UUCP already working -- just fill in the screen form telling it how to dial and login to the local backbone site. I spent one Sunday bringing up netnews, starting with the PD source distribution. (Now that I've done it, if any other 6300PLUS-users want it, it's available here, free, any time!) Definitely worth it! -- Dave Levenson Westmark, Inc. A node for news. Warren, NJ USA {rutgers | clyde | mtune | ihnp4}!westmark!dave
bds@lzaz.ATT.COM (BRUCE SZABLAK) (09/16/87)
UNIX(TM) has the demonstrated ability of running in a variety of hardware environments: micro to mainframe. Given recent history: 8008->8080->8086-> 80186->80286->80386 and CPM->MSDOS->OS/2; it seems that this is essential for any competing OS. The question is: can OS/2 be true to its traditions, still have a future when the 80386 is history, and not collapse under its own weight in the end? My personal opinion is there is no good reason to take that risk; UNIX(TM) will be there on the next generation of processors, and the generation following that one...
feg@clyde.UUCP (09/16/87)
In article <498@parcvax.Xerox.COM>, burton@parcvax.Xerox.COM (Philip M. Burton) writes: > > And if I don't want to do it, and I'm faily technical for a non-developer > (at least the guys at work tell me that), then it's a hopeless situation > for the person who buys a PC just to do spreadsheets of maintain a database. > This average Joe user that you describe---afraid to install Wordstar, or is frightened of backups, and does a little with 123----why would this user ever want a 386? What would he do with it? ANY multiuser, multi- tasking OS is going to seem like a labyrinth to oblivion to this person. On another part of this subject: it has taken Microsoft 5 years to file off the edges of ms-dos to a point that it is a fairly reasonable OS for a personal computer. If you have been following the predictions for OS/2 availability--you have seen it slipping off into the distance. Even when that great day arrives, and you have it up and running on your 386, it will be a great disappointment to many people. My guess is that about 1992 it will start to be more like what many expect it to be next year. Forrest Gehrke
daw@houxs.UUCP (Dave Wolverton) (09/16/87)
In article <965@looking.UUCP>, brad@looking.UUCP (Brad Templeton) writes: [ lots of good stuff deleted ] > > The Unix command structure and administration structure does need to > change. As I noted before, the Unix hacks are the block to this. > Right now Unix sells mainly to people already sold on Unix, or people > who buy it as part of an OEM package. It loses others because of the > reputation Unix has. If you change the commands etc. you lose the old > folks and wait a long time to get the new folks. Not profitable. > Perhaps most people are unaware that the UNIX boxes which AT&T sells (3B2/600, 3B4000, etc.) all come with a menu-oriented program called "sysadm" which provides a naive-user level interface for 90% of the administration tasks. For instance, I can format a floppy, make a file system on it, and mount the file system, all inside sysadm. It also provides a mechanism for automated installation of applications. In our workgroup, we have computer-naive clerks perform much of the administration of our minis using this interface. Also, someone in this discussion mentioned that the popularity of the Macintosh indicated that people don't want command-line interfaces. I tend to disagree with this conclusion. The Mac folks indicate that what they really want is the ability to open a UNIX window *when they need it*, but otherwise use the Mac OS more-or-less as it is. Also, note that the Amiga, which also uses a windowing interface, has always had a command line interface (CLI) window for those who want it. > Unix is a trade mark of Bell Labs. Actually, the correct incantation is "UNIX is a registered trademark of AT&T." At least you got the period in the right place :-) :-) :-) > -- > Brad Templeton, Looking Glass Software Ltd. - Waterloo, Ontario 519/884-7473 Dave Wolverton AT&T Somethingorother Labs ihnp4!houxs!daw (Warning: I don't usually read this newsgroup)
sl@van-bc.UUCP (09/17/87)
In article <498@parcvax.Xerox.COM> burton@parcvax.xerox.com.UUCP (Philip M. Burton) writes: >In article <961@looking.UUCP> brad@looking.UUCP (Brad Templeton) writes: >However, menus are part of the answer. For many people, any kind of command >line is NOT the answer. That's a good part of the success of the Mac. Menus may not be the answer, but the current System V Sysadm package is at least adequate for reducing the amount of knowledge you need about all of the nitty gritty details to a low level. It basically steps you through all of the steps necessary to add users, System file entries (aka uucp L.sys file), etc. I know how to do all this, but still use Sysadm, because it keeps me from forgetting things. >> The biggest problem is the difficulty in administering and maintaining a >> Unix system. It is still orders of magnitude more difficult than the >> "turn it on and run a program" simplicity of DOS. Many of these problems >> are not specific to Unix, but are caused by having dynamic, multi-tasking, >> multi-user networked environments. OS/2 will fall plague to some of these. Exactly. Anytime you add a hard disk and five or ten MB of software together you start having problems. I for one don't have *any* problems bringing up a Unix system. However I cringe anytime I have to go near an MS-DOS system. It's really not very easy to setup. I mean really, auto-exec what?? For me to bring 3.2 MS-DOS on our recently installed 386 would take me longer than it did to put unix on it (about 1hr from a standing start). The reason that everyone thinks that MS-DOS is easier is that they have had a relatively slow and long training period, with lots of peer support. Everyone had MS-DOS so it was easy to find other people who could help you. And the real experts started with 1.0 (ah, a great little version, the last one *I used*), and progressed upwards through levels two, and on to 3.[012]. For people like me who havn't kept up it's really quite annoying to try and do anything with it. I know all of the basic ideas but don't know any of the specific little details. And when I compare the basic capabilities of MS-DOS to Unix, I usually just give up. And if you've ever brought up an IBM pc network, you can't really claim that MS-DOS is a simple system to configure! Makes *anything* else look simple. Basically the only MS-DOS that is easy to configure is two floppy system with 256k RAM :-) >At the same time, until you make fundamental, and I do mean that, changes >to UNIX, for example the filesystem fragility, lack of true random access, I've beening running on an older Uniplus Unix box for over a year now, with over 150MB of files on second hand, slightly flakey hard disks. The system crashes occasionally (flakey hardware). I've *never* lost a file. With the current release of System V, release 3, AT&T has "hardened" the file system. Reports of the "fragility" of the Unix file system are grossly mis-informed. >you don't have a PERSONAL operating system. Funny, I'm the only PERSON on this system. Of course I've got two terminals and two phone lines plugged into it, but I'm the only one who uses it as their main system. I do allow about twenty other computers to route mail and news through it, otherwise it would just sit there and get dusty. It might as well be doing *something*. >> The big advantage Unix has is that Unix for the 80386 is available now. And cheap too! I heartily recommend it. >> OS/2 for the 386 is predicted by Microsoft to be over a year away. If And probably not so cheap too! >> the delivery is anything like that for MS Windows, well.... And probably it will be about as successfull as MS Windows too! >No, that conclusion about UNIX doesn't follow. UNIX will not run software >written for OS/2. And even if it were true, my "marketing gut" tells me that Say's who. If you think that the people who brought you DOS/Merge / Locus etc, are not carefully studing the OS/2 manuals right now you are just naive. I would venture to say that you may be able to run 386 OS/2 programs under Unix before you can under OS/2! >What got me started on all this was the prospect that MicroPort has a UNIX >Then I started thinking of the incredible hassles that I would have to put >up with to get uucp news up, or whatever. Just not worth it. Unix is available for very low prices on the 386. The 386 is the first machine that will allow large scale distribution of binary compatible Unix programs over a wide range of different manufacturers machines. I for one am currently porting news and a large number of PD tools over to my machine at work. Once I've done this I'll be able to hand out copies of the binaries to anyone interested. You want news "cpio < /dev/rdsk/f0q15dt". I think that Unix will replace MS-DOS wherever users want multi-user or multi-tasking. It's not all that hard to use/install/configure. You just have to learn how to do it, just like you did with MS-DOS. -- {ihnp4!alberta!ubc-vision,uunet}!van-bc!Stuart.Lynne Vancouver,BC,604-937-7532
jgray@toad.UUCP (09/17/87)
In article <680@hsi.UUCP>, tankus@hsi.UUCP (Ed Tankus) writes: > > Phillip Barton made some interesting comments. His point of view is from the > marketing end rather than from the developer/hacker/guru perspective. Each > has a point to be made and each has done it and done it well. > > Now the bottom line is this: OS/2, U*IX, or ABCD operating system will NOT > take over the world. Over time, each will have its own market, and hopefully, > enough overlap, that we will start moving towards a more homogenous environment. > ....... > > The arguing can go on, but lets not continue. I think most will agree that > we have aired enough of our own views to last a few months until the next > discussion on this same subject starts anew. > > Thanks to all those who did have something to say. It was very interesting. > Now let's move on. > > > Cheers! > > -- Ed. > Actually I've been enjoying the discussion so far, but Ed makes a good point. As far as which OS is better in the market place, let the subject rest for a while. I personally would like to hear more technical comparisons between OS/2 and Unix so that I can make an educated choice. By technical, I mean something that a software/hardware developer could appriciate. For example could someone explain what they mean by a Unix file system being more vulnerable than DOS? Or is the Unix v.s. OS/2 comparison equally valid for 286 and 386 based machine? And so on..... --------------- Jerrold L. Gray UUCP:{ihnp4|caip|tektronix|ucbvax}!uw-beaver!tikal!pilchuck!jgray USNAIL: 10525 Willows Road N.E. /C-46 Redmond, Wa. 98052 (206) 881 - 6444 x470 Telex: 15-2167
burton@parcvax.UUCP (09/18/87)
In article <13909@clyde.ATT.COM> feg@clyde.ATT.COM (Forrest Gehrke) writes: >In article <498@parcvax.Xerox.COM>, burton@parcvax.Xerox.COM (Philip M. Burton) writes: >> >This average Joe user that you describe---afraid to install Wordstar, or >is frightened of backups, and does a little with 123----why would this >user ever want a 386? What would he do with it? ANY multiuser, multi- >tasking OS is going to seem like a labyrinth to oblivion to this person. > Perhaps this average Joe does big, big spreadsheets all day. Now, on his PC, this takes 25 minutes to do a re-calc. On his AT, maybe 5 minutes. But, on his Compaq 386, less than 2 minutes. (These are real numbers, drawn from the experiences of a major San Francisco bank.) Now, if this average Joe can get a 386 that runs DOS, that's great. Apparently, there are lots of Joes out there, judging by Compaq's sales of the 386. -- Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton Xerox Corp. preferred path: burton.osbunorth@xerox.COM 408 737 4635 ... usual disclaimers apply ...
mitchell@cadovax.UUCP (Mitchell Lerner) (09/18/87)
On the subject of Unix's command interface being more user friendly...
I've recently seen two product that are called somthing like Directory
Monitors or editors or somthing like that. What they seem to do is
provide a windowed "point and go" approach to the Unix commands.
It's alot like some most of the tools that I've seen for DOS: One scrolls
through the possible commands with the space bar or arrow keys; has a
choice of editors to use (one even had a HEX editor which you could
patch disks and files with).
It seemed very comprehensive and easy to use and it had the ability to
generate new windows for work areas and alot of things. I guess the point
is that a few people have already created solutions to the Unix environment
being for technical types only.
--
Mitchell Lerner
UUCP: {ucbvax,ihnp4,decvax}!trwrb!cadovax!mitchell
"When I fight with my mind, my mind always wins" - Ben Hummelmerchant@dartvax.UUCP (09/18/87)
In article <13909@clyde.ATT.COM>, Forrest Gehrke writes: > This average Joe user that you describe---afraid to install Wordstar, or > is frightened of backups, and does a little with 123----why would this > user ever want a 386? What would he do with it? ANY multiuser, multi- > tasking OS is going to seem like a labyrinth to oblivion to this person. A couple of reasons. One, he's the head of the company, so he wants the power. Of course, he's also probably a Bozo and we don't much care about him. The second reason is one that I've tried to explain to people around here. The argument goes that why do students need 80286 personal computers? 8086/8088 based systems are certainly powerful enough for the things they'll have to do. Personally, I've been saying they should get 80286 or 80386 based machines. But that's just me. Why do these clueless people need 80386 power? Because they are clueless! If I don't know what I'm doing, I need a computer system that has a very simple to understand interface. Probably graphical. That requires power. If I'm a lousy speller, I need spelling and grammar checkers. Those require power. In essence, our clueless user needs power because the computer has to do more work to make him more productive. Isn't that why we use our personal computers? -- "Anyway you want it, Peter Merchant (merchant@dartvax.UUCP) That's the way you need it."
c60b-ia@buddy.Berkeley.EDU (Sugih Jamin) (09/19/87)
In article <2944@hoptoad.uucp> farren@hoptoad.UUCP (Mike Farren) writes: >In article <494@parcvax.Xerox.COM> burton@parcvax.Xerox.COM writes: >> >>Now, for the average user who is terrified of installing Wordstar or backing >>up files, or who uses a hard disk without any subdirectories [too much trouble >>to get them to work right], let's tell that person about UNIX. Any version. >> >No, you don't tell them any of those things. You set up a system, as a >shell, which is as easy to use as MS-DOS (and probably a lot easier), but >which also allows access to deeper, more cryptic levels of Unix if they >want them. You provide, within that system, the tools which are necessary >to let them do their work, and you make sure that the tools are easy-to-use. >You provide, within your Unix implementation, as much protection as possible >against the types of file corruption that you cite (it isn't easy, but it's >possible). Then, you let them go. > >Unix, in its history, has been viewed as an OS for the computer sophisticate. >In part, this is true, in that to use all of its capabilities requires a >fairly large amount of knowledge of its structure. The "normal" user, >however, doesn't need to do this type of access at all, and Unix is very >nicely set up to allow differing types of access for differing types of >needs. All it takes is a little work, and, unfortunately, few people seem >to be very interested in doing this work. "Well, *I* understand it, so >why can't they?" seems to be a prevalent attitude, and it is WRONG. > I would like to make some comments on the last sentence quoted above. But let me comment on the desirability of writing a "protective" shell for UNIX first. OK, so I'll setup UNIX to be as simple as one of the simpler OS. I'll make it such that my users don't have to know anything about login, logout, chmod, ln, etc. (I am not sure the first two are possible in a multi-users system.) I'll make it such that "a:" will bring them to the floppy; such that everything the users want to do will be questioned, as is the prevalent philosophy in one of the "icon-based" OS. I'll go to the trouble of writing tons of shell scripts, aliases, or even little utilities, debug them, safeguard them, maintain them. (And all those after paying for a UNIX box which, so far, costs much more than either of the other OS mentioned above.) All for the purporse that my users won't be able to play with the power of UNIX and get burned. Why bother? Why not, instead, get one of the cheaper, simpler, already "working" system that I am supposed to "emulate"? Maybe my users will one day want to use the full power of UNIX? For that kind of users, I'll advice learning the real personality of UNIX as early as possible, and they are not our main point of discussion. We are talking about people who put everything in their home directory and never use any sub- directory because "it doesn't worth the bother to get it working right" as farren@hoptoad put it. (I don't remember the exact wording.) Even burton@parcvax himself agreed that we are talking about "the 'normal' users who don't need to do this type of access at all." For these users, I say, they want a PC or a Mac, so give them what they want; and I certainly am not going to disguise and downgrade UNIX to "emulate" either. So, you see, my main reason for not making some stupid "menus" is not "Well, *I* understand it, why can't they?" But because I believe it is a better solution for those kind of user to get the kind of machine they really need in the first place. >We can, and should, make it possible to use Unix as easily as using MS-DOS, >or even as easily as using the Mac. Alias manipulation alone will, I think, give an almost-MS-DOS-like shell. Our kind of users, who turn on their machine just to get a letter written and printed out, presumably will know only the commands erase, type, ren, md, cd, rd, print, and maybe mv, plus ws, and lotus, and dBaseIII. As for Mac emulation, you certainly need a mouse and a comparable screen resolution, how about a SUN or uVax running X-window? Somebody over here is working on a "Finder's look-alike" for the X-window. (Don't ask, I am not involve with that project and know nothing more about it; as I am more interested in getting X-window running on the Mac with a UNIX-like shell--too bad RASNIX was such an incomplete attempt.) sugih jamin (c60b-ia.berkeley.edu)
jay@splut.UUCP (Jay Maynard) (09/20/87)
In article <17@lzaz.ATT.COM>, bds@lzaz.ATT.COM (BRUCE SZABLAK) writes: > UNIX(TM) has the demonstrated ability of running in a variety of hardware > environments: micro to mainframe. Given recent history: 8008->8080->8086-> > 80186->80286->80386 and CPM->MSDOS->OS/2; it seems that this is essential > for any competing OS. The question is: can OS/2 be true to its traditions, > still have a future when the 80386 is history, and not collapse under its > own weight in the end? > > My personal opinion is there is no good reason to take that risk; > UNIX(TM) will be there on the next generation of processors, and > the generation following that one... Except for one minor detail: Un*x is NOT the portable solution everyone claims it is. If it were, we wouldn't have all these discussions about "package X wants to run on System V, can someone help me to get it going on 4.3BSD?" I've had little success taking software off the net and making it go. Maybe POSIX (or one of the other standards efforts) will succeed in this regard, but I doubt it. -- Jay Maynard, K5ZC (@WB5BBW)...>splut!< | uucp: hoptoad!academ!uhnix1!splut!jay Never ascribe to malice that which can | or sun!housun!nuchat!--^ be adequately explained by stupidity. | GEnie: JAYMAYNARD CI$: 71036,1603 The opinions herein are shared by neither of my cats, much less anyone else.
acm@bu-cs.UUCP (09/20/87)
In article <224@trwrc.UUCP> agnew@trwrc.UUCP (R.A. Agnew) writes: >>benefits. (The marketplace isn't a few engineers here and there, but the >>millions and millions of white collar workers who drive cars with automatic >>transmissions and never built a Heathkit in their lives and use cameras with >>auto-exposure and automatic flash.) >> >I couldn't agree more! Would all of you millions of white collar workers >go off and play with your 123 and OS-2 and leave Unix where it belongs. >After all, it was written by engineers, for engineers. (Note the >bessel and error functions in the C run time library and the plot >and graph routines ). Meanwhile, I'll design your next user friendly >PC for you. I'll make sure not to make it too hard to use. What bessel and error functions in the C run time library? Having a pretty good notion of the history of C/UNIX, I think you're confused. First, are Kernigan/Ritchie engineers? Perhaps software engineers. Second, each C library has the routines supplied by the writer. The libraries are not standard even among UNIX's, much less among non-UNIX machines. I also find it unlikely that many C's have plot functions since that is quite hardware-dependant! Sorry, flame that's somewhat off the subject. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Jim Frost * The Madd Hacker | UUCP: ..!harvard!bu-cs!bucsb!madd H H | ARPA: madd@bucsb.bu.edu H-C-C-OH <- heehee +---------+---------------------------------- H H | "We are strangers in a world we never made"
acm@bu-cs.UUCP (09/20/87)
In article <1344@van-bc.UUCP> sl@van-bc.UUCP (Stuart Lynne) writes: >In article <498@parcvax.Xerox.COM> burton@parcvax.xerox.com.UUCP (Philip M. Burton) writes: >[...] the current System V Sysadm package is at >least adequate for reducing the amount of knowledge you need about all of >the nitty gritty details to a low level. It basically steps you through all >of the steps necessary to add users, System file entries (aka uucp L.sys >file), etc. I know how to do all this, but still use Sysadm, because it >keeps me from forgetting things. Having used the AT&T system, let me tell you it's pretty easy to install things, get them running, add users, perform system checks, etc. A bit slow, perhaps, but very very easy. The only thing that is a pain to DOS users is the floppy disk -- you CANNOT remove it without telling the system. Does anyone out there own a Mac? What happens when you use a paper clip to remove a disk while inside an application? Sometimes, nothing. Other times, boom. Hmm -- people can handle macs, even though you have to tell the system to give you your disk back. Methinks they could handle UNIX if you give 'em a reasonable interface. >[...] I cringe anytime I have to go near an MS-DOS system. >It's really not very easy to setup. I mean really, auto-exec what?? For me >to bring 3.2 MS-DOS on our recently installed 386 would take me longer than >it did to put unix on it (about 1hr from a standing start). It shouldn't take very long to bring up anything but the most complex DOS system. I can bring up just about any DOS system in less than an hour and usually in a few minutes, even with multiple hard drives. UNIX is harder to bring up, but (again) it's more complex. >The reason that everyone thinks that MS-DOS is easier is that they have had >a relatively slow and long training period, with lots of peer support. Everyone >had MS-DOS so it was easy to find other people who could help you. And the >real experts started with 1.0 (ah, a great little version, the last one *I >used*), and progressed upwards through levels two, and on to 3.[012]. Nah. The reason they think it's easy is because it is -- it's pretty braindamaged so of course it's simpler. You introduce multitasking and multiuser and you immediately make life tougher on the people who set up and use the system. >And if you've ever brought up an IBM pc network, you can't really claim that >MS-DOS is a simple system to configure! Makes *anything* else look simple. True! Again, multiuser.... The only real problem with DOS here is that it was not designed for multiuser access to the filesystem. >>At the same time, until you make fundamental, and I do mean that, changes >>to UNIX, for example the filesystem fragility, lack of true random access, > >I've beening running on an older Uniplus Unix box for over a year now, with >over 150MB of files on second hand, slightly flakey hard disks. The system >crashes occasionally (flakey hardware). I've *never* lost a file. Mr. Burton: WHAT filesystem fragility? Compared to MS-DOS? Are you serious? You really have to break something to lose information on a UNIX box! Compare this to MS-DOS, where the only way to allocate bad sectors (aside from a third-party utility) is to reformat your hard disk. Where a single bad sector in a directory can trash an entire directory substructure with absolutely no way of recovering your valuable data. Also, I notice that there was a hell of a market for Norton's utility package, which should indicate that users have problems with MS-DOS. The UNIX file system is pretty nice and quite consistant. It's easy to find (and fix) errors within it. DOS's is not. DOS's filesystem is also grossly inefficient -- when you move up to larger hard disks, you end up with allocation clusters of 2K or more and no way to pack smaller files together. The UNIX fragmentation method is much more efficient. Also, I notice that DOS has a 32Mb limit on lots of things. Want to check out the size limitation on UNIX? Can you say "unlimited"? Granted, files become less efficient (slightly) when you approach the gigabyte range, but you have to be reasonable. (I'm leaving out a lot of information on how UNIX does its file storage but if you really want to see what I mean you could buy a book on the internals of UNIX.) Don't knock the UNIX filesystem until you've tried it, and don't compare it to MS-DOS's. MS-DOS's structure is horrible compared to UNIX's. >With the current release of System V, release 3, AT&T has "hardened" the file >system. Reports of the "fragility" of the Unix file system are grossly >mis-informed. True. And even the old one was more robust than MS-DOS's. >>> OS/2 for the 386 is predicted by Microsoft to be over a year away. If > >And probably not so cheap too! Definitely not so cheap. IBM's price (straight from their catalogue) is $395/copy. And it's still vaporware.... This brings up a question. Anyone know how many people are paying for copies of OS/2 now? It *was* advertised in IBM's catalogue (with "To be released first quarter 1988" in very small print), as was the model 80 and some enhancements to OS/2 (including the Presentation Manager which should be out sometime before the turn of the century). I'd like to know what the sales are for OS/2 before release. >>What got me started on all this was the prospect that MicroPort has a UNIX > >>Then I started thinking of the incredible hassles that I would have to put >>up with to get uucp news up, or whatever. Just not worth it. Heehee. You think it's a hassle to get uucp news up under UNIX? Wanna try it with DOS? Hmm? Do you? *I* don't! >I think that Unix will replace MS-DOS wherever users want multi-user or >multi-tasking. It's not all that hard to use/install/configure. You just >have to learn how to do it, just like you did with MS-DOS. It's already beginning to do that in many companies. One of the favorite schemes I've seen is to use a UNIX system as a multi-PC fileserver to PC's running MS-DOS. This gives the security and reliability of the UNIX filesystem to MS-DOS while allowing users to run MS-DOS and its software. On a final note, it would be naive to believe that OS/2 will fall prey to UNIX. Very, very naive. One thing that I've learned since being in business is that people will wait for IBM even if the wait is much longer than expected, even if there are better products available, even if the product is less than advertised, even if the product is overpriced, even if the product is faulty. There are numerous examples I can pull up to back up this claim (many in my own experience). I won't include them here unless someone asks. OS/2 or UNIX? I'd take UNIX. But I'm a programmer. Ask DP managers and I bet you most of them will answer OS/2. They're drooling for it, even though better is available right now. If you have comments, please feel free to give them. I love an argument! Besides, I don't know everything yet.... %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Jim Frost * The Madd Hacker | UUCP: ..!harvard!bu-cs!bucsb!madd H H | ARPA: madd@bucsb.bu.edu H-C-C-OH <- heehee +---------+---------------------------------- H H | "We are strangers in a world we never made"
bcw@rti.UUCP (Bruce Wright) (09/21/87)
In article <13212@bu-cs.BU.EDU>, acm@bu-cs.BU.EDU (ACM) writes: > > Definitely not so cheap. IBM's price (straight from their catalogue) > is $395/copy. And it's still vaporware.... This brings up a > question. Anyone know how many people are paying for copies of OS/2 > now? It *was* advertised in IBM's catalogue (with "To be released > first quarter 1988" in very small print), as was the model 80 and some > enhancements to OS/2 (including the Presentation Manager which should > be out sometime before the turn of the century). I'd like to know > what the sales are for OS/2 before release. It is not at all clear to me that _either_ OS/2 or Unix will amount to a hill of beans in the PC marketplace outside of niche markets. Before you flame me, hear me out. I define things like LAN servers, high-end CAD workstations, medical imaging systems, high-end desktop publishing, etc. as niche markets. For low-end CAD, or low-end desktop publishing, or the vast majority of applications used by PC users, it just doesn't provide any significant benefit. Multi- tasking is nice, but it's really most useful to support multi-user applications (hence LANs) or specialized applications requiring lots of horsepower and lots of user interaction *at the same time*. The typical spreadsheet, word processor, terminal emulation program, or what have you, just doesn't have any _need_ for multitasking - it's not even in its domain. The only thing about OS/2 which is at all interesting to the majority of users will be the graphical interface - what IBM calls the Presentation Manager. This will not be released for who knows how long - certainly _long_ after the base release of OS/2. And there are plenty of graphics interfaces now (Windows, GEM, etc) which are not exactly setting the world on fire - there are some people who like them but there is hardly a wholesale migration to any of them. About the only benefit that most PC users would see from improved multi- tasking is a better print spooler. Period. And better print spoolers are available as add-on programs or as cheap print buffer hardware. A better printer spooler could be put on MS-DOS - the main problem with the one there now isn't that you can't do print spooling reasonably on a system without complete multi-tasking (see RT-11, which does have a Foreground/ Background but which are hardly a complete tasking environment), but that the parallel interface on the IBM monochrome card is brain-damaged such that you can't run it with interrupts - so MS-DOS doesn't even try with other interfaces since this one is so likely. You could fix this problem with PS/2 or other hardware and better support from MS-DOS or with a third- party printer driver. Add to this the fact that OS/2 will only run on the 286 and 386 machines and you have a real niche market for OS/2 (and Unix). Of course the major developers will support OS/2 - why shouldn't they??? If you are running a LAN or a CAD workstation which might sometimes use (for example) 1-2-3, you _don't_ want to hear that the program won't run on that OS. But this won't change the niche market for OS/2 - since the developers will in all likelihood not want to cross off the 88/86/186 markets - especially since the low-end PS/2 machines will still be running the older chips. If it were cheap, or if there were some really useful, widely-needed product which would run on OS/2 but not on MS-DOS, OS/2 (or Unix) might become as widely used as MS-DOS. But I'm not holding my breath. Bruce C. Wright
feg@clyde.UUCP (09/21/87)
In article <514@parcvax.Xerox.COM>, burton@parcvax.Xerox.COM (Philip M. Burton) writes: > In article <13909@clyde.ATT.COM> feg@clyde.ATT.COM (Forrest Gehrke) writes: [Why would "average Joe" Burton describes want a 386?" > >In article <498@parcvax.Xerox.COM>, burton@parcvax.Xerox.COM (Philip M. Burton) writes: > >> > > [deletions] > > Perhaps this average Joe does big, big spreadsheets all day. Now, on his > PC, this takes 25 minutes to do a re-calc. On his AT, maybe 5 minutes. But, > on his Compaq 386, less than 2 minutes. (These are real numbers, drawn > from the experiences of a major San Francisco bank.) > > You are undoubtedly correct. But your original subject was directed at the reasons OS-2 was going to be better than UNIX as seen by your average Joe user. When he is using ms-dos on a Compaq 386, he is not seeing any difference from his experience with the earlier PC's except speed. The introduction of multi-tasking and multi-user, (even if multiple users are not present) capability to an OS, any OS, will make OS-2 look a lot like UNIX. Even ms-dos after v 2.0 has many obvious borrowings from UNIX. I predict that OS-2 will look even more like UNIX, regardless of what it is called and regardless of what the commands are called (erase/delete instead of rm, and rename instead of mv, for example). Unix development began in 1969. It is still evolving, rel 3 now beginning to show up for Sys V. Since ms-dos has taken 5 years to get to v 3.3 (with most of the bugs out and showing reasonable features for a single task/single user OS), I don't think 5 years is going to be an unreasonable wait for a decently appearing OS-2 (even with UNIX as a template before the developers). And it is going to be a great deal more complicated to install and run than ms-dos. Worse yet, OS-2 on your Compaq 386, is going to run more slowly than ms-dos does. Forrest Gehrke
root@hobbes.UUCP (John Plocher) (09/22/87)
+---- lots of DOS/Unix "talk" deleted ----+ The biggest difference I've found in the DOS -vs- Unix war is that I can set up a development environment (WP, C, debugging...) on a 10Mb hard disk under DOS and still have about 8 Meg free for my programs. A similarly configured Unix system wouldn't even fit on 10Mb! Why should Joe Dosuser bother with "disk hog" Unix? What can Unix do that is worth burning all that disk space for? In the user's perspective there is little (read NO) difference between WordStar under DOS and ditroff under Unix - they both produce "word processing" output. The same thing goes for Turbo C -vs- pcc/lint/dbx and Brief -vs- emacs. Besides, Sidekick and Superkey and Procomm and Lotus 123 don't work under Unix. 1/2 :-( On Unix, things can change without your doing anything: Log files get bigger, spool directories fill up, etc. Sure, it's a result of multi-tasking, but it confuses the hell out of people who panic when they see two copies of their report on a directory listing (never mind the fact that one of them ends in ".BAK" :-) Don't get me wrong, I use Unix (with DosMerge) because I know what it can do for my productivity, but I don't expect ANY of the users I deal with to come anywhere near the level I need to be at to keep the system running! Jane Novice can keep her DOS machine going with NO maintenance. NONE. Come in in the morning, turn it on; work on files during the day; copy everything to a floppy (copy file.123 a:); and turn it off before going home. Most PC users don't even bother with copying to a floppy :-(. Try that with Unix! I find myself spending about 20 minutes a day maintaining each of out Unix boxes PLUS 1/2 an hour extra on Fridays doing backups. I shudder to think what things would be like if I took off for 2 weeks and no one was here to look after things! -- John Plocher uwvax!geowhiz!uwspan!plocher plocher%uwspan.UUCP@uwvax.CS.WISC.EDU
randy@chinet.UUCP (09/24/87)
In article <153@splut.UUCP> jay@splut.UUCP (Jay Maynard) writes: >In article <17@lzaz.ATT.COM>, bds@lzaz.ATT.COM (BRUCE SZABLAK) writes: >> UNIX(TM) has the demonstrated ability of running in a variety of hardware >> environments: micro to mainframe. > >Except for one minor detail: > >Un*x is NOT the portable solution everyone claims it is. If it were, we >wouldn't have all these discussions about "package X wants to run on System >V, can someone help me to get it going on 4.3BSD?" I've had little success >taking software off the net and making it go. Hmmm.. Let's see.. I have this fairly large, integrated software package, let's call it news 2.11. I, personally, have access to: VAX 8600 running SySV2, 3b2, 3b1, AT running XENIX, Microport, Pyramid 90x running 4.2bsd and SysV, Microvax II running ULTRIX. I typed "make" on all of them, and was running a complicated communications system in no time. I have LOTUS 123 running on my AT. Now, how bout you take the source and get that running as fast (or within a couple of magnitudes of fast) on my Mac. Methinks UN*X has *something* going for it.. -- that's the biz, sweetheart..... Randy Suess ..!ihnp4!chinet!randy
richardh@killer.UUCP (09/24/87)
In article <230@hobbes.UUCP>, root@hobbes.UUCP (John Plocher) writes: > > The biggest difference I've found in the DOS -vs- Unix war is that I can set > up a development environment (WP, C, debugging...) on a 10Mb hard disk under > DOS and still have about 8 Meg free for my programs. A similarly configured > Unix system wouldn't even fit on 10Mb! Why should Joe Dosuser bother Are you implying that what is true for DOS will also be true for OS/2? > with "disk hog" Unix? What can Unix do that is worth burning all that > disk space for? > For one thing, try comparing the Unix shell sh to DOS's "batch" "language" ?! There are many Unix programmers who never get past the shell. They don't have to ! And many user's benefit from this capability. > Besides, Sidekick and Superkey and Procomm and Lotus 123 don't work under > Unix. 1/2 :-( > Are you implying that they will under OS/2? > Jane Novice can keep her DOS machine going with NO maintenance. NONE. > Come in in the morning, turn it on; work on files during the day; copy > everything to a floppy (copy file.123 a:); and turn it off before going home. > Most PC users don't even bother with copying to a floppy :-(. > I think Jane may be unpleasantly surprised by OS/2. > -- > John Plocher uwvax!geowhiz!uwspan!plocher plocher%uwspan.UUCP@uwvax.CS.WISC.EDU Actually, I agree with you in your DOS vs. Unix comparison. I think a lot of people are under-estimating the staying power of DOS. Because it is so simple to use (and most things run faster under DOS - there are no annoying kernel tasks running along with yours, no mailer, no sync daemon, no spooler, etc.), I think many users will say "OS/2 (or Unix for that matter)? Just say no." Running DOS and Unix on the same system is very educational; I heartily recommend it to anyone who wants to get a real feel for the differences. regards, richard hargrove ...!ihnp4!killer!richardh -------------------------
mike@cimcor.UUCP (Michael Grenier) (09/24/87)
In article <230@hobbes.UUCP>, root@hobbes.UUCP (John Plocher) writes: > +---- lots of DOS/Unix "talk" deleted ----+ > > The biggest difference I've found in the DOS -vs- Unix war is that I can set > up a development environment (WP, C, debugging...) on a 10Mb hard disk under > DOS and still have about 8 Meg free for my programs. A similarly configured > Unix system wouldn't even fit on 10Mb! Why should Joe Dosuser bother > with "disk hog" Unix? Actually, a simialarly configured Unix will fit in the space you mention - My Microport boot disk provides most or all of the things that DOS does. Sure its single user mode, and doesn't have the print spoolers, and the uucp communications, etc - but neither does DOS! { I don't consider PRINT under dos much of a spooler}. One usually doesn't configure UNIX to run in that small of space because UNIX gives you so much more that its WORTH making space for it and most people do. Concerning John's other claim in his article about the need for UNIX system administration and cleaning up log files - again the above argument I make holds : If you don't install mail, UUCP, spoolers, etc like DOS then you don't have to maintain them either. -Mike {ihnp4,rutgers,amdahl}!meccts!cimcor!mike
kerr@oodis01.UUCP (09/24/87)
Is it true that microsoft uses OS/2 on their sun workstations or that apple uses OS/2 on their cray? (or was it unix?) :-) -- Grant Kerr Control Data Corporation INTERNET: kerr@oodis01.arpa UUCP: ihnp4!lll-tis!oodis01!kerr
jay@splut.UUCP (Jay Maynard) (09/24/87)
In article <13212@bu-cs.BU.EDU>, acm@bu-cs.BU.EDU (ACM) writes: > In article <1344@van-bc.UUCP> sl@van-bc.UUCP (Stuart Lynne) writes: > >In article <498@parcvax.Xerox.COM> burton@parcvax.xerox.com.UUCP (Philip M. Burton) writes: > >[...] I cringe anytime I have to go near an MS-DOS system. > >It's really not very easy to setup. I mean really, auto-exec what?? For me > >to bring 3.2 MS-DOS on our recently installed 386 would take me longer than > >it did to put unix on it (about 1hr from a standing start). You mean it takes an hour for you to stuff a disk in A: and turn the machine on? > It shouldn't take very long to bring up anything but the most complex > DOS system. I can bring up just about any DOS system in less than an > hour and usually in a few minutes, even with multiple hard drives. > UNIX is harder to bring up, but (again) it's more complex. Boy, you can say that again. I've had to start over and reinstall my Microport system three times, and there are still a couple of installation decisions I need to change - which means a fourth re-installation... > >>At the same time, until you make fundamental, and I do mean that, changes > >>to UNIX, for example the filesystem fragility, lack of true random access, > > > >I've beening running on an older Uniplus Unix box for over a year now, with > >over 150MB of files on second hand, slightly flakey hard disks. The system > >crashes occasionally (flakey hardware). I've *never* lost a file. > > Mr. Burton: > > WHAT filesystem fragility? Compared to MS-DOS? Are you serious? You > really have to break something to lose information on a UNIX box! Yeah...your finger reaching for the power switch. (only 1/2 :-)...as I sit here and type with an injured finger...) > Compare this to MS-DOS, where the only way to allocate bad sectors > (aside from a third-party utility) is to reformat your hard disk. I have a bad sector now on my second hard drive, and I'll have to reformat it to lock it out - and it's exclusively for Unix (worse - it's exclusively for news! :-) > Where a single bad sector in a directory can trash an entire directory > substructure with absolutely no way of recovering your valuable data. A single bad sector in the root directory of a filesystem will do the same thing... > Also, I notice that there was a hell of a market for Norton's utility > package, which should indicate that users have problems with MS-DOS. Most of which was so that people could (more-or-less easily) unerase files. > The UNIX file system is pretty nice and quite consistant. It's > easy to find (and fix) errors within it. Only if you trust fsck - and I'm not sure I do, especially given the comment in the Unix manual (I don't recall exactly where) that recommends that you have file system expertise available for fixing broken filesystems... > Don't knock the UNIX filesystem until you've > tried it, and don't compare it to MS-DOS's. MS-DOS's structure is > horrible compared to UNIX's. At lease I can fly 'chkdsk -f' and understand what's going on. Anyone want to explain fsck in less than 10 pages? > >With the current release of System V, release 3, AT&T has "hardened" the file > >system. Reports of the "fragility" of the Unix file system are grossly > >mis-informed. > > True. And even the old one was more robust than MS-DOS's. Not on my system, it isn't... > >>> OS/2 for the 386 is predicted by Microsoft to be over a year away. If > > > >And probably not so cheap too! > > Definitely not so cheap. IBM's price (straight from their catalogue) > is $395/copy. And it's still vaporware.... Who said Unix is cheap?? $450 for System V/AT isn't peanuts...and that's the buggier of the various Unixes available for the PC - but it also happens to be the cheapest. > OS/2 or UNIX? I'd take UNIX. But I'm a programmer. Ask DP managers ...or users... > and I bet you most of them will answer OS/2. They're drooling for it, > even though better is available right now. Better by whose definition? If you think in terms of a strictly program development environment, then you're probably right in thinking of Unix - for it's a superior system for people who are willing to take the time and effort to learn incredibly cryptic commands, multiuser system administration (even on boxes that only get used by one person), and babying a complex and somewhat fragile system. If you think in terms of a user who's only interested in turning on the computer and getting an answer on how many widgets he needs to buy supplies to make next year, then Unix is a terrible non-solution. > If you have comments, please feel free to give them. I love an > argument! Besides, I don't know everything yet.... So do I, and I certainly don't know all about Unix by any stretch of the imagination. (Just ask Peter Da Silva, with whom I've had this argument on and off for four years...) -- Jay Maynard, K5ZC (@WB5BBW)...>splut!< | uucp: hoptoad!academ!uhnix1!splut!jay Never ascribe to malice that which can | or {sun!housun,uunet}!nuchat!-^ be adequately explained by stupidity. | GEnie: JAYMAYNARD CI$: 71036,1603 The opinions herein are shared by neither of my cats, much less anyone else.
braun@drivax.UUCP (09/24/87)
In article <1729@rti.UUCP> bcw@rti.UUCP (Bruce Wright) writes: >It is not at all clear to me that _either_ OS/2 or Unix will amount to >a hill of beans in the PC marketplace outside of niche markets... >Multitasking is nice, but it's really most useful to support multi-user >applications (hence LANs) or specialized applications requiring lots of >horsepower and lots of user interaction *at the same time*. The typical >spreadsheet, word processor, terminal emulation program, or what have you, >just doesn't have any _need_ for multitasking - it's not even in its domain. This is the Jerry Pernoulle (sp?) perspective and I think it is just flag *wrong*. A glance at the plethora of ram resident software (and programs to (control all of these) is proof. More than just for print spooling, time sharing is needed for *any* background task: modem control, keyboard mapping, etc. At least suspended-task technology is required for the other things. How many times have you been in your spread sheet and needed to look up something in another file? How many times have you been editing your source file and needed to look up the spelling of some constant in an include file? Ok, windowing will address a lot of this, whether it is in the application (the editor) or the graphics system (GEM, Windows, etc). But this type of resource management belongs with the rest of the resource management: In the OS, not in the application level. -- kral 408/647-6112 ...{ism780|amdahl}!drivax!braun "Dream lightyears... Challenge miles... Walk in steps" DISCLAIMER: If DRI knew I was saying this stuff, they would shut me d~-~oxx
sl@van-bc.UUCP (Stuart Lynne) (09/25/87)
In article <13212@bu-cs.BU.EDU> madd@bucsb.bu.edu writes: >In article <1344@van-bc.UUCP> sl@van-bc.UUCP (Stuart Lynne) writes: >>In article <498@parcvax.Xerox.COM> burton@parcvax.xerox.com.UUCP (Philip M. Burton) writes: >>[...] I cringe anytime I have to go near an MS-DOS system. >>It's really not very easy to setup. I mean really, auto-exec what?? For me >>to bring 3.2 MS-DOS on our recently installed 386 would take me longer than >>it did to put unix on it (about 1hr from a standing start). > >It shouldn't take very long to bring up anything but the most complex >DOS system. I can bring up just about any DOS system in less than an >hour and usually in a few minutes, even with multiple hard drives. >UNIX is harder to bring up, but (again) it's more complex. My point is (was) that I've brought Unix on numerous small boxes and don't have any trouble doing so. I've had my learning curve and know basically what to do. On the other hand I've never used MS-DOS on a hard disk (I'll admit to 1.1 on floppies), at least not for anything other than editing a few files. So if I had to bring up an MS-DOS system I wouldn't know where to begin. Even the few times I've just tried to make a simple floppy based MS-DOS system I end up extremely frustrated from the cruddy tools and documentation. The first time you bring up any system its hard, Unix *is* probably harder. But once you learn either of them they arn't that hard to work with. Although going from a powerful system like Unix to MS-DOS can be frustrating due to the lack of the same tools. -- {ihnp4!alberta!ubc-vision,uunet}!van-bc!Stuart.Lynne Vancouver,BC,604-937-7532
greg@xios.XIOS.UUCP (Greg Franks) (09/25/87)
In article <497@parcvax.Xerox.COM> burton@parcvax.xerox.com.UUCP (Philip M. Burton) writes: >Another fundamental problem was highlighted by Bill Gates at a recent >meeting of the Silicon Valley Computer Society recently. UNIX has a model of >dumb terminals communicating via serial lines to a host. DOS and OS/2 assume >a much tighter linkage. > >How many serial line-connected UNIX machines actually support bit-mapped >graphics? Actually, SUN has an answer. Get yourself an EL-CHEAPO Atari ST. Toss in some custom software - namely a postscript display processor. (Toss out GEM etc). Connect it to your SUN and voila. I don't even think you need anything more elaborate than an RS232 line. I haven't seen one personally :-( - I would like to give one a whirl! See comp.sys.atari.st for other bits of trivia. -- Greg Franks XIOS Systems Corporation, 1600 Carling Avenue, (613) 725-5411 Ottawa, Ontario, Canada, K1Z 8R8 uunet!mnetor!dciem!nrcaer!xios!greg "Vermont ain't flat!"
pat@seradg.Dayton.NCR.COM (Patrick Pesch) (09/26/87)
In article <476@cimcor.UUCP> mike@cimcor.UUCP (Michael Grenier) writes: >In article <230@hobbes.UUCP>, root@hobbes.UUCP (John Plocher) writes: >> +---- lots of DOS/Unix "talk" deleted ----+ >> >> The biggest difference I've found in the DOS -vs- Unix war is that I can set >> up a development environment (WP, C, debugging...) on a 10Mb hard disk under >> DOS and still have about 8 Meg free for my programs. A similarly configured >> Unix system wouldn't even fit on 10Mb! Why should Joe Dosuser bother >> with "disk hog" Unix? > >Actually, a simialarly configured Unix will fit in the space you >mention - My Microport boot disk provides most or all of the things >that DOS does. Sure its single user mode, and doesn't have the print >spoolers, and the uucp communications, etc - but neither does DOS! [...] Hurrah! Somebody has finally made the distinction between a high power monitor program [DOS :-)] and a full blown operating system complete with useful utilities, development tools, etc... Also, comparing the costs of each of the above configurations for Joe User can be suprising... Unix packages include "C" compilers, development tools, text processing, choice of user shells (SCO Xenix has 3), communications, electronic mail, print spooling, and (most) have some games to boot! Let's make sure we are comparing "operating systems" or included utilities, etc... when we compare Unix/ OS/2 / DOS. *** *** ******* ********* **** *** ********* *********** Patrick Pesch ***** *** *** ** *** **** NCR Corporation ****** *** *** *** *** Systems Engineering Retail *** ****** *** ********* Dayton OH *** ***** *** ** *********** *** **** ********* *** **** pat@seradg.Dayton.NCR.COM *** *** ******* *** **** "Standard Employee Disclaimer Applies" - Plato
dhesi@bsu-cs.UUCP (09/27/87)
In article <1611@killer.UUCP> richardh@killer.UUCP (Richard Hargrove) writes: >Are you implying that what is true for DOS will also be true for OS/2? MS-DOS, CP/M, OS/2, UNIX, AmigaDOS, TRSDOS, are all DOSes. If something is true of every DOS it's true of OS/2. -- Rahul Dhesi UUCP: <backbones>!{iuvax,pur-ee,uunet}!bsu-cs!dhesi
cy@ashtate (Cy Shuster) (09/28/87)
Gordon Lettwin, the chief architect for OS/2 at Microsoft, characterized the
difference between OS/2 and Unix (at the OS/2 Developer's Conference) as the
difference between multitasking, single user and multitasking, multi-user.
Unix was designed as a multiuser timesharing system, and its design goal was
to maximize throughput on the shared resources: disk, memory, and printers.
OS/2, in contrast, has been design to maximize *response* for the single user
controlling multiple tasks. The paradigm for an OS/2 application is for the
process monitoring user input to create a new thread to handle a users'
request, while the main thread immediately waits for further user input.
An example of how these different design goals can result in different
system behavior is that under OS/2, when the system detects an overload
condition (e.g., memory thrashing), it can send an alert box to the (single)
user, informing him of the resource overload, and then asking him which
tasks to cancel to alleviate the situation -- since all tasks were started
by the same user. This would obviously not pertain to the multiuser situation!
Gordon freely acknowledges that MS-DOS is barely a program loader, never mind
an operating system. Almost all of the sucessful MS-DOS programs to date have
been those that cleverly manage to work around the performance restrictions
intrinsic to DOS, by staying resident, writing directly to video memory, and so
on. This has made hackers out of the lot of us, forced to write highly hardware
and software-dependent code. With OS/2 and its dynamic linked libraries, we can
begin to actually use the operating system to do low-level work, and the end
user can run software from different companies without worrying about RAM cram
and which TSR to load last.
It seems certain that OS/2 acceptance is still a ways off, however. The native
8086 emulation that Microsoft persuaded Intel to put into the 386 chip is
about to spawn a host of 386 control software like DeskView 2.0, allowing
multiple DOS 3 sessions on a single machine, without requiring the end user to
buy all new software.
The end user still doesn't care what the operating system is, if he is suffi-
ciently insulated from it. The current topic of file system fragility is very
germane. I, for one, favor the Lisa approach, where hitting the big red switch
to power off merely generates a shutdown request to the operating system, which
proceeds to neatly terminate active tasks, and save off the current desktop
state before signalling a solenoid to pull the switch. I can't believe how many
users do the "three-finger crash" as the normal way to exit software!
--Cy Shuster UUCP: ...seismo!scgvaxd!ashtate!cy
DISCLAIMER: Do you have any *idea* what I would have to go through to get
this approved as an official opinion???? MINE ONLY!bcw@rti.UUCP (Bruce Wright) (09/28/87)
In article <2440@drivax.UUCP>, braun@drivax.UUCP (Kral) writes: > ... More than just for print spooling, time > sharing is needed for *any* background task: modem control, keyboard mapping, > etc. At least suspended-task technology is required for the other things. How > many times have you been in your spread sheet and needed to look up something > in another file? How many times have you been editing your source file and > needed to look up the spelling of some constant in an include file? > The above quote was in response to my criticism of OS/2 and Un*x as an operating system for the masses. Unfortunately I think it proves my point. These tasks are most often done by techies whose environment would be more accurately described as a workstation and hence would be a logical OS/2 or Un*x target. But the people who form the bulk of the PC marketplace don't do a lot of this - they have enough trouble with disks, files, and so forth. We may not like it, but that's the way it is - these people are pretty confused already, and will not want to go to the added confusion of a complex, expensive and (to them) incomprehensible multitasking system. Brucit iences.en
burton@parcvax.Xerox.COM (Philip M. Burton) (09/28/87)
In article <1740@rti.UUCP> bcw@rti.UUCP (Bruce Wright) writes: >In article <2440@drivax.UUCP>, braun@drivax.UUCP (Kral) writes: >> ... More than just for print spooling, time >> sharing is needed for *any* background task: modem control, keyboard mapping, >> etc. At least suspended-task technology is required for the other things. How >> many times have you been in your spread sheet and needed to look up something >> in another file? How many times have you been editing your source file an Couple of issues here. First, you are mixing user types when you mention spreadsheets and source files. Had you said correspondence, you would habve been on target. Real users do however, print big spreadsheets, and Lotus, as an example, just freezes up untuil the print job is over (or at least spooled to the printer's buffer). It's easy for the average Epson printer to take five minutes on a big sheet. Sure would be nice the I could save or change m y file while the printer is banging away. Or start my letter. Or, if my big spreadsheet takes 10 minutes to load and do all initial calculations, it would be nhice if I could get that done as a background task while I am typing my memo. Or recalcing while printing my memo. Or even formatting my disks. Or maybe, just maybe, I want to run TWO Lotus speradsheets at once, either in windows or on separate virtual screens. Or whatever. The notion that only us techies, read high priests of computerdo, need a multi-tasking system is pure unadulterated horse feathers. (It's also a way not to get rich. -)). (mailer wasted space) -- Philip Burton burton@parcvax.COM ...!hplabs!parcvax!burton Xerox Corp. preferred path: burton.osbunorth@xerox.COM 408 737 4635 ... usual disclaimers apply ...
rcw@qetzal.UUCP (09/29/87)
In article <161@splut.UUCP>, jay@splut.UUCP (Jay Maynard) writes: > > I have a bad sector now on my second hard drive, and I'll have to reformat > it to lock it out - and it's exclusively for Unix (worse - it's exclusively > for news! :-) Jay refers to Microport SysV/AT. It is not necessary to reformat the disk. Simply :-) use the fdisk utility, the "scan and assign" bad tracks option to fix a sector on the fly. I agree that badsector handling is clumsy under unix, and different on every unix machine I've seen. How about a consistent badsector handling interface for the user? > A single bad sector in the root directory of a filesystem will do the same > thing... [trash the file system] Or, God forbid, the free list. > At lease I can fly 'chkdsk -f' and .... Ever try to use one of the files that this command recovers? > [Unix] Better by whose definition? If you think in terms of a strictly program > development environment, then you're probably right in thinking of Unix - > for it's a superior system for people who are willing to take the time and > effort to learn incredibly cryptic commands, multiuser system administration It'd be nice to close the case on this meta-dos vs. unix discussion. I think that most would agree that DOS is suitable for single-user applications like Joe Blow's financial database, and that Unix is more appropriate for multi-user applications/software development/ graphic workstations/University environments. The key difference is *multi-user*, and one would be naive to think that a multi-user os is going to be trivial to maintain. (They are also naive to even *think* about purchasing OS/2 before release 3.2) -- Robert C. White, Jr. Mentor Software, Inc. UUCP: ihnp4!upba!qetzal!rcw isis!qetzal!rcw USPS: 3067 Robin Way, Denver, CO 80222 ATT : +1 303 759-3666
jgray@toad.pilchuck.Data-IO.COM (Jerry Late Nite Gray) (09/29/87)
In article <1740@rti.UUCP>, bcw@rti.UUCP (Bruce Wright) writes: > In article <2440@drivax.UUCP>, braun@drivax.UUCP (Kral) writes: > > ... More than just for print spooling, time > > sharing is needed for *any* background task: modem control, keyboard mapping, > > etc. At least suspended-task technology is required for the other things. How > > many times have you been in your spread sheet and needed to look up something > > in another file? How many times have you been editing your source file and > > needed to look up the spelling of some constant in an include file? > > > > The above quote was in response to my criticism of OS/2 and Un*x as an > operating system for the masses. Unfortunately I think it proves my point. > > These tasks are most often done by techies whose environment would be more > accurately described as a workstation and hence would be a logical OS/2 > or Un*x target. But the people who form the bulk of the PC marketplace don't > do a lot of this - they have enough trouble with disks, files, and so > forth. We may not like it, but that's the way it is - these people are > pretty confused already, and will not want to go to the added confusion of a > complex, expensive and (to them) incomprehensible multitasking system. > > Bruce C. Wright Sorry but this isn't what I've been observing. I haven't seen a non-techie here at this company who doesn't know(and doesn't use frequently) the feature that pressing both shift keys of their PC will suspend VTERM so they can look at, or execute, other things. Maybe we are unusual. Task suspension is a very simple concept that non-techies are already using and each tool implements differently (Procomm uses Alt-F4, my Emacs uses Esc-@ and so on..). Unix tools all suspend with Control-Z. I can't for the life of me understand where having one method for suspension for all tools (which OS/2 should have) would be more complicated. Take a more complicated system feature like remote access to your PC across a network. The best that has been done in a DOS-like environment is simple file serving. In our NFS network of non-PCs I can "login" to any of the machines on the network and run programs there interactively or remotely in addition to simple file serving. Trying to get the same thing in a collective DOS environment would be impossible if not impractical since the operating system is simply inadequate. Inadequate to a degree that probably could not be made up for by a plethora of TSRs, drivers and "cleverly" written application programs. The point made previously, and stressed here, is that the operating system is the best place for certain types of solutions and would simplify the environment. Multitasking eliminates the need for TSRs. Multi-user would make more "interesting" features happen in a PC of which there is little or no equivalent "kludges" under DOS. I don't believe that the non-existence of these kludges under DOS implies "no market". Some things are just too difficult to be practical under DOS. If you think that a multi-user environment equates only to multiple people using the same hardware, you are only partially correct. It also means multiple people sharing resources with an inherent mechanism in the operating system to support I.D's, file protection schemes (can you say "chmod"?) and the like. If you look at a network oriented DOS compatible operating system called Qnix (I'm not an advocate of this system) you'll see what I mean. --------------- Jerrold L. Gray UUCP:{ihnp4|caip|tektronix|ucbvax}!uw-beaver!tikal!pilchuck!jgray USNAIL: 10525 Willows Road N.E. /C-46 Redmond, Wa. 98052 (206) 881 - 6444 x470 Telex: 15-2167
agnew@trwrc.UUCP (R.A. Agnew) (09/30/87)
In article <13211@bu-cs.BU.EDU> madd@bucsb.bu.edu (Jim "Jack" Frost) writes: > >What bessel and error functions in the C run time library? Having a >pretty good notion of the history of C/UNIX, I think you're confused. I think you're confused. There's not a Bell systems programmer's manual out there without j0(), jn(), graph, and plot since before Version 7. Better check again!
hoff@hp-sdd.UUCP (09/30/87)
Anyone out there using Microport System 5 (or whatever they call it) on an
AT class machine? I have some questions like:
Can you share files with DOS if you boot the machine up in DOS?
How much disk space does the OS and all the support utilities take?
Does it have all the nice stuff like vi, awk, cc, etc?
How does it compare to other *ix systems like SCO or Microsoft XENIX?
Are device drivers availiable for popular hardware accessories?
How much does it cost?
Where can I get a free copy of all source code? :-)
Please send email!!!!! IF I get a good response I'll post a summary.
Much Grass,
--Tom
--
Tom Hoff (...!hplabs!hp-sdd!hoff)
"Dammit Jim, I'm a programmer not a spokesman!"tankus@hsi.UUCP (10/01/87)
[Text removed for brevity] > > It'd be nice to close the case on this meta-dos vs. unix discussion. > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [Ditto] > > -- > Robert C. White, Jr. Mentor Software, Inc. > UUCP: ihnp4!upba!qetzal!rcw isis!qetzal!rcw > USPS: 3067 Robin Way, Denver, CO 80222 > ATT : +1 303 759-3666 I THIRD Robert White's second of my original motion of a week (or more) ago to end this discussion! I think most will agree everyone has played their hand. Let's clear the airwaves. For those who wish to continue, please use your mailers. THANKS! -- Ed. Net : {uunet,ihnp4,noao,hao,yale}!hsi!tankus Snail: Health Systems Int'l, 100 Broadway, New Haven, CT 06511 Bell : (203) 562-2101
ralf@b.gp.cs.cmu.edu (Ralf Brown) (10/03/87)
In article <177@qetzal.UUCP> rcw@qetzal.UUCP (sysop) writes: >In article <161@splut.UUCP>, jay@splut.UUCP (Jay Maynard) writes: >*multi-user*, and one would be naive to think that a multi-user >os is going to be trivial to maintain. (They are also naive to >even *think* about purchasing OS/2 before release 3.2) ^^^ are you sure ? Remember how buggy DOS 3.2 was? And 3.3.... (and MASM 4.0, and and and...) >Robert C. White, Jr. Mentor Software, Inc. >UUCP: ihnp4!upba!qetzal!rcw isis!qetzal!rcw >USPS: 3067 Robin Way, Denver, CO 80222 >ATT : +1 303 759-3666 DOS 3.1 forever! -- -=-=-=-=-=-=-=-= {harvard,uunet,ucbvax}!b.gp.cs.cmu.edu!ralf =-=-=-=-=-=-=-=- ARPAnet: RALF@B.GP.CS.CMU.EDU BITnet: RALF%B.GP.CS.CMU.EDU@CMUCCVMA AT&Tnet: (412) 268-3053 (school) FIDOnet: Ralf Brown at 129/31 DISCLAIMER? Who ever said I claimed anything? "I do not fear computers. I fear the lack of them..." -- Isaac Asimov
bcw@rti.UUCP (Bruce Wright) (10/05/87)
In article <711@pilchuck.Data-IO.COM>, jgray@toad.pilchuck.Data-IO.COM (Jerry Late Nite Gray) writes: > > Sorry but this isn't what I've been observing. I haven't seen a non-techie > here at this company who doesn't know(and doesn't use frequently) the feature > that pressing both shift keys of their PC will suspend VTERM so they can > look at, or execute, other things. Maybe we are unusual. I think you _are_ unusual. I suspect that the reason is that there appears to be a fair amount of interaction between your technical and non-technical users; in most of the environments I've seen this is not the case. > Task suspension is a very simple concept that non-techies are already using And task suspension isn't the same as multi-tasking. Anyway, see above. > Take a more complicated system feature like remote access to your PC across > a network. The best that has been done in a DOS-like environment is simple > file serving. This is at least debatable. In any event "logging in" to a remote system has value only if the remote system has something the local system doesn't - such as power or information. PC's are sufficiently similar in terms of power that this is not a major issue - the spread of speed between the fastest and slowest is only about 10x, and if a fast PC is sharing its time between several users over some sort of comm lines, it's not obvious that it's going to be faster than a slow local machine. Sharing of information can be done with file servers; it's not immediately obvious how this is a significant multi-tasking issue (not that you imply it is). In any event I have doubts about how easy it would be for _any_ non-technical user to set up any sort of network environment that was much more than just a plug- in-the-box item, and I have some reservations about that. I speak with some experience because I am involved with the development and support of network products on non-PC hardware and am constantly amazed with how little the _technical_ users know about what they are using, let alone the _non- technical_ people. > The point made previously, and stressed here, is that the operating system > is the best place for certain types of solutions and would simplify the > environment. No question about it. But that's _NOT_ the issue!!! Users (and especially non-technical users) aren't particularly interested in how many "kludges" there are in a product if it (at least sort of) fills their perceived needs. And I've _NEVER_ said that there would be "no market" (YOUR words) for OS/2. I _HAVE_ said that it would be a "niche" market. There's a _VAST_ difference. I have also never said that you couldn't have a pretty large "niche" market - just that it would by no means rival the size of the MS-DOS market. When OS/2 has been out long enough to have things everyone seems to take for granted like the Windows interface, and for some applications to be ported to it, there are likely to be something on the order of 15-20,000,000 PC-type machines out there (there's already on the order of 10M). Many of them would be _incapable_ of running OS/2 because of CPU or memory (RAM or disk) limitations; many of the remaining ones would be running MS-DOS out of basic human conservatism. It's optimistic to think that there would be more than 1M OS/2 installations at that point - significant but a clear minority. It's also not by any means clear that this state of affairs won't continue for the forseeable future. > It also means multiple people sharing resources with an inherent > mechanism in the operating system to support I.D's, file protection schemes > (can you say "chmod"?) and the like. I find this insulting. In addition I don't think I have heard of a very complete file protection scheme for OS/2, and it's not so obvious what that means (other than protection against accidental deletion) on a single-user but multi-tasking system. Judging from some of the network traffic, some people seem to have the idea that I am dead set against OS/2 or any multi-tasking system. This is not the case - I am simply pointing out that the market at this point is _very_ conservative. There have been other reasonable multi-tasking systems for the PC (Unix and Concurrent DOS, for example) which have failed miserably in the basic PC market outside of certain niches. OS/2 may find itself a larger niche, but the market has already spoken - unless multi-tasking is either _very_ cheap (not much more than, say, $50 over a single-tasking system) or _necessary_ for some critical application (say 1-2-3 or something similar), it appears unlikely _in the extreme_ for it to become a _majority_ system on the PC for a very long time (if ever). This is especially true if the system is not _exactly_ compatible with MS-DOS. I'm not at all certain I _like_ this state of affairs, but that's how it seems to be. And I don't think OS/2 should be ignored - in fact there are some products which I am contemplating for targeting to OS/2 - but I don't think that it can be regarded as the de facto PC operating system. That's all. Bruce C. Wright
acm@bu-cs.BU.EDU (ACM) (10/08/87)
In article <177@qetzal.UUCP> rcw@qetzal.UUCP (sysop) writes: >In article <161@splut.UUCP>, jay@splut.UUCP (Jay Maynard) writes: >> A single bad sector in the root directory of a filesystem will do the same >> thing... [trash the file system] Yea, but it'll do it for MS-DOS too. In UNIX the root directory doesn't HAVE to have anything in it (except perhaps boot image) -- just mount file systems on it. This is what I would do. So you get a bad sector -- kill that cylinder and use the next one up. You lose a few K, is all, and restoring data is trivial because there wasn't anything there. Now, MS-DOS also needs only the boot image there, but a bad sector could screw up the directory links. This is a problem since the directories aren't mounted, they are on that physical device. This is sort of a problem, wouldn't you say? I'd still take the UNIX method over MS-DOS's. >> At lease I can fly 'chkdsk -f' and .... > >Ever try to use one of the files that this command recovers? He has a point. I *have* tried to use 'em. Not too easy, let me tell you. Also, chkdsk -f won't fix all kinds of things. If the FAT gets destroyed, you're in trouble. I don't think chkdsk will even attempt to rebuild the entire FAT; I know that fsck DOES, and does it with some redundancy. >The key difference is >*multi-user*, and one would be naive to think that a multi-user >os is going to be trivial to maintain. (They are also naive to >even *think* about purchasing OS/2 before release 3.2) True. And, true. But they'll never get 3.2 out until they get 1.0! %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Jim Frost * The Madd Hacker | UUCP: ..!harvard!bu-cs!bucsb!madd H H | ARPA: madd@bucsb.bu.edu H-C-C-OH <- heehee +---------+---------------------------------- H H | "We are strangers in a world we never made"