wayne@ames.arc.nasa.gov (Wayne Hathaway) (08/11/88)
My six cents worth (three two-centers), all relating to UNIX-WIZARDS Digest V6#004: First, ok@quintus.uucp (Richard A. O'Keefe) quotes and adds: >> In article <670025@hpclscu.HP.COM> shankar@hpclscu.HP.COM (Shankar Unni) writes: >> :How do you think IBM sells anything? Product quality is usually priority >> :number 1 (like the Ford commercial :-)) at most of the big players in the >> :business. > Have you ever used VM/CMS? Or its Pascal compiler? Please note that the original quote went on to say >> Hell hath no fury like the DP manager whose end-of-the-month >> payroll run has been disrupted by a system failure. I personally know of no DP managers who run payrolls under VM/CMS, especially not in Pascal. The point that Shankar was making is that IBM became IBM by selling an industrial-strength product that did what it was supposed to do, period. [And by the way, for what it's worth VM/CMS is still out there more in SPITE of IBM than because of it.] Next ditto@cbmvax.UUCP (Michael "Ford" Ditto) adds: > Yes, we all know this based on past experience. Remember the IBM PC > (naww, hardly ever see them anymore :-)? IBM entered into a field it > had no experience in, took some off-the-shelf parts using 5-year-old > technology, bought an operating system from a company mainly selling > BASIC interpreters, called it a product, threw some marketing budget > at it, and look what happened. From MY experience, what IBM did was collect together a bunch of toy components and a toy operating system and beat the hell out of it for several months (many tens of man-years worth, if memory serves) bringing it up to "IBM quality" (whatever that really was) so that it wouldn't break or crash or corrupt data every few minutes. As someone who was actually making a living on pre-IBM PCs under CP/M (does the name Micromation ring any bells any more?), I for one think what IBM did really WAS the start of the "PC Revolution," at least the commercial side of it (I'll leave the other side to the Steves et al). Again, the same point: industrial strength quality, something a business (or a manager) could afford to risk the future on. [I am reminded of a comment from an IBM officer at a SHARE meeting many moons ago, objecting to people saying that IBM made "the Cadillacs of computers" -- his thesis was that IBM made TRUCKS, not Cadillacs.] Then madd@bu-cs.BU.EDU (Jim Frost) added: > This is partially right. IBM became big by being reliable; they never > did anything really new so what they had was most likely going to > work. With the 360 and 370 series machines they got people locked > into an architecture; which was the better thing to do when your > system was too small? Buy a nice, new, fast machine (real cheap) or > stick with IBM and not have to re-code anything? Less headaches with > IBM, so they stayed. Did I miss something, or isn't this the same discussion that includes a thread about how the SVID can be (or can't be, depending upon the discussant :-) capriciously changed, requiring massive software rewrites to catch up? If so, it sure seems strange to see IBM chastised for recognizing (25 years ago, for crying out loud!) that software porting (upgrading, re-coding, whatever) was going to be so expensive that they could make a mint by selling less cost-effective hardware that avoided it. [I also participated in a "migration" from a 7094 (good old IBSYS, for the other Neanderthals in the audience :-) to a 360, and believe me, I'd happily pay a LOT more for hardware to avoid doing THAT again!] Finally, lest anybody think I am particularly pro-IBM or anti-UNIX or anything, let me add that the best thing I've read in this whole discussion was somebody's line that "IBM supports UNIX like a rope supports a hanging man." Cheers! Wayne Hathaway ultra!wayne@Ames.ARC.NASA.GOV Ultra Network Technologies 2140 Bering drive with a domain server: San Jose, CA 95131 wayne@Ultra.COM 408-922-0100
peter@ficc.UUCP (Peter da Silva) (08/16/88)
In article ... ultra!wayne@ames.arc.nasa.gov (Wayne Hathaway) writes: > From MY experience, what IBM did was collect together a bunch of toy > components and a toy operating system and beat the hell out of it ...so that it... > wouldn't break or crash or corrupt data every few minutes. Are we talking about the same piece of junk MS-DOS 1.0 that I remember? About the only advantage was that the stupid error message was changed from "BDOS ERROR ON C: SELECT" to "ABORT, RETRY, IGNORE". In either case it meant "YOU JUST LOST YOUR DATA, SUCKER". > As someone who was actually making a living on pre-IBM PCs under CP/M... And we're talking about a different CP/M, too. I don't recall any serious deficiencies in CP/M that weren't shared by the early PC-DOS... There were some crummy CP/M boxes, yes, but there were good ones as well. -- Peter da Silva, Ferranti International Controls Corporation, sugar!ficc!peter. "You made a TIME MACHINE out of a VOLKSWAGEN BEETLE?" "Well, I couldn't afford another deLorean." "But how do you ever get it up to 88 miles per hour????"
dhesi@bsu-cs.UUCP (Rahul Dhesi) (08/16/88)
In article <1260@ficc.UUCP> peter@ficc.UUCP (Peter da Silva) writes: >And we're talking about a different CP/M, too. I don't recall any serious >deficiencies in CP/M that weren't shared by the early PC-DOS... CP/M had no expandability. With MS-DOS you *began* with 64 K, remember? With CP/M, you finished there. With the IBM PC you got a turnkey machine with slots for expandability, a standard BASIC-in-ROM, and the promise of repair service *nationwide*. To the best of my knowledge there was no other machine in the same price range at that time that had this. You'll be surprised how many people (perhaps close to 100% of the first users) wrote all their applications in BASIC. With competing CP/M machines, there was no standard programming environment, since everybody's video display and language interpreter was slightly different. The IBM PC was the only machine for which you could write a good graphics program (for the color/graphics display) and expect it to work on all machines. But CP/M did last a little longer, thrashing and desperately struggling to survive, while MS-DOS added a hierarchical file system, loadable device drivers, I/O redirection and pseudo-pipes, and the ability to know how long a file was, not just guess. And when we went shopping, we found that we didn't have to spend $150 for CP/M-86, since MS-DOS was only forty bucks! And then, when Lotus 1-2-3 came out, and it worked only on an IBM PC using MS-DOS, the final blow had been struck against CP/M. The DEC Rainbow was the only other system that could boast of a nationwide service network, but what with the lack of software and hardware expandability and the horrendously unreliable disk drives, it was doomed from the beginning. Then DEC further confused its users by providing both CP/M and MS-DOS, but carefully making sure you could neither transfer files between the two nor format disks under either. Copyright 1988 Rahul Dhesi. Permission granted for Usenet distribution. -- Rahul Dhesi UUCP: <backbones>!{iuvax,pur-ee,uunet}!bsu-cs!dhesi
ram%shukra@Sun.COM (Renu Raman) (08/17/88)
In article <3660@bsu-cs.UUCP> dhesi@bsu-cs.UUCP (Rahul Dhesi) writes: >In article <1260@ficc.UUCP> peter@ficc.UUCP (Peter da Silva) writes: >>And we're talking about a different CP/M, too. I don't recall any serious >>deficiencies in CP/M that weren't shared by the early PC-DOS... > >CP/M had no expandability. With MS-DOS you *began* with 64 K, >remember? With CP/M, you finished there. > If you want more than 64K, there is a way. You can have banks of 64K memory and switch between them. My Visual 1050 has 128K (2 banks) + 32K for display and I believe, I can add another 128K to it. >-- >Rahul Dhesi UUCP: <backbones>!{iuvax,pur-ee,uunet}!bsu-cs!dhesi
mlinar@eve.usc.edu (Mitch Mlinar) (08/17/88)
In article <3660@bsu-cs.UUCP> dhesi@bsu-cs.UUCP (Rahul Dhesi) writes: > >CP/M had no expandability. With MS-DOS you *began* with 64 K, >remember? With CP/M, you finished there. > Unless you had banked memory CP/M. However, it is still true that any given task could only be performed in 64k. My "typical" CP/M setup these days has 1M of RAM, multitasking, and lots of RAM disk at 10MHz of 64180 power. It blows the doors the normal PC and fights for speed with the original AT. Although it is clearly not as powerful as an AT by most measures, as matched to an IBM PC, there are only two differences . Graphics was *the* key, no doubt about it. Extra usable program memory certainly helped, but it also made some programs turn into RAM pigs. Given the inefficiency of IBM compilers as compared to the tight code you *had* to have for CP/M, the ratio is somewhere between 2:1 or 3:1 in code size. The same C program (compiled on a number of compilers like MS-DOS, Lattice, and Power C) by myself [to prove a point in Feb of this year], averaged 2.8x larger on the IBM than CP/M. Yes, the code should be bigger since the average instruction length is longer, but not by that much. Only the data structures are the same size. Of course, if you throw in lots of floating pt, and/or you happen to have a co-processor, this ratio drops slightly. > >And then, when Lotus 1-2-3 came out, and it worked only on an IBM PC >using MS-DOS, the final blow had been struck against CP/M. > which is why MicroPro is selling WordStar4 for CP/M (released just this past year) and MicroSoft was considering re-releasing some of their 8-bit stuff!?!!? In fact, I have seen my software sales INCREASE on CP/M over the past 18 months. I would *hate* to see what my sales base would be if CP/M was not dead as you claimed .... :-) :-) I am now going to twist this around on you: IBM DOS is dead. I think that CP/M and early IBM DOSes are converging on the same obsolete boat. With (a) UN*X, (b) OS/2, (c) no future versions of DOS, (d) no future machines with anything less than a 80286 and (e) major vendors jumping to the newer OSes, IBM DOS must be dead [paraphrased from 3 computer journals during the past months]; this sounds awful familiar to me. And, just like CP/M, the death of IBM DOS will be announced every year for the next decade because it refuses to *really* die. Companies are moving away from IBM-PC 8088 for 386/Suns/MacIIs, but a large chuck of home computers will still be around. Now and ten years from now. The "appliance" users *always* move on as they can afford it. But if you have a computer at home, you are more than likely a hobbyist without the bank in your hip pocket to buy the latest machine that comes along every couple of years. I have been through it already with CP/M; IBM-PC friends who are sweating OS/2s impact on the little 8088 machine need not worry. IBM DOS will not vaporize overnight either. As long as the computer meets your needs, whatever those are, and there are others who feel the same, it is not obsolete. -Mitch