rmilner@zia.aoc.nrao.edu (Ruth Milner) (06/05/91)
Here's a nice general topic for discussion. It occurred to me the other day that there was a UNIX consortium, made up of various computer vendors, which was founded either shortly before or shortly after the OSF. I vaguely recall (but this may be cynical hindsight :-) ), that right after the first one was formed, everyone who hadn't joined it rushed out and founded the second one. What was that other consortium, and why don't I hear anything about it any more? Has it simply changed its name or something? (Did I just dream it up? I was out of things for several months between jobs, maybe it disbanded then). Does anyone else feel as skeptical as I do about all these groups of companies getting together to co-operate on things? Will it really all work out nicely, or will we just trade in the BSD/SysV dichotomy for SVR4/OSF? It's nice to see them collaborating on benchmarks, to keep each other honest, but I can't help thinking that the notion of everyone truly winding up with the same UNIX blend is just pie in the sky. And then there's the IEEE, the OSI, and ANSI sticking their noses in, too ... Who was it who said "the nice thing about standards is that there are so many of them"? -- Ruth Milner Systems Manager NRAO/VLA Socorro NM Computing Division Head rmilner@zia.aoc.nrao.edu
rcd@ico.isc.com (Dick Dunn) (06/06/91)
rmilner@zia.aoc.nrao.edu (Ruth Milner) writes: > Does anyone else feel as skeptical as I do about all these groups of > companies getting together to co-operate on things?... No...I feel much more skeptical than you do. >...Will it really all work out nicely, > or will we just trade in the BSD/SysV dichotomy for SVR4/OSF?... The dichotomies will continue. (OK, what's the word? "polychotomy"? or just "confused mess"?) New variants of "UNIX" (using the term *very* loosely) are appearing faster than the unification efforts are moving. Moreover, each (re)unification adds a great pile of sh^H^Hentropy--we get lots of new code to make the compatibility happen, but we get no new capabilities, and performance is no better, probably worse. The consortia serve a more nefarious purpose, in my opinion. Suppose one vendor tells you, "you just gotta have this here-now X window system and this GUI and window manager and desktop manager...and oh, by the way, you're going to have to add 8 meg of memory, get another disk, and upgrade the CPU so that performance will be adequate." You'd tell that vendor where to put his gui little window system, and look elsewhere. Now try the same scenario with a mega-consortium. They're going to tell you what's good for you, and leave you as few choices about where else to turn as possible. You might get away with questioning the direction of a DEC, a Compaq, or an HPollo...but anyone who questions a consortium of a dozen or so respected vendors is clearly a fringe case. They want to set up the situation so that they can tell you what you have to buy, and limit your choices to the things they want to build--things which will sell hardware, and allow "upgrades" that keep selling more and more hardware, mostly for trivial cosmetic changes--because substantive changes cost real money to make. The great corporate discovery of the late '80's to the unpredictability and expense of constructing significant software (with bloated, mismanaged groups of mediocre people) is that you can avoid the unpredictability by not doing anything significant. If you can't make software good enough to be sold on its merits, just make something--anything--and go wild on marketing to sell it regardless of its merits. Consortia provide the means to market this software kitsch, to drown out the voices of reason crying "who wants this crap?" (Or maybe these are just the rantings of a paranoid who's heard one too many conspiracy theories...but it does seem curious that the consortia are top- heavy with hardware vendors trying to dictate software that requires lots of expensive new hardware.) [If you're expecting a disclaimer after that, you don't understand USENET.] -- Dick Dunn rcd@ico.isc.com -or- ico!rcd Boulder, CO (303)449-2870 ...Simpler is better.