Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!crdgw1!crdos1!davidsen
From: davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr)
Newsgroups: comp.arch
Subject: Re: Wet Dream or ...
Message-ID: <3358@crdos1.crd.ge.COM>
Date: 19 Apr 91 14:59:20 GMT
References: <13964@adobe.UUCP> <1991Apr11.184136.10324@news.iastate.edu> <1991Apr12.151301.27159@pcserver2.naitc.com>
Reply-To: davidsen@crdos1.crd.ge.com (bill davidsen)
Distribution: na
Organization: GE Corp R&D Center, Schenectady NY
Lines: 70

In article <1991Apr12.151301.27159@pcserver2.naitc.com> kdenning@pcserver2.naitc.com (Karl Denninger) writes:

| I have reasonably reliable information that SCO Unix will be dead as of
| December of this year.  It's being replaced.
| 
| The replacement?  OSF/1!  

  Do you remember the customer reaction when SCO tried to jump everyone
from Xenix to UNIX? If they repeat the process, particularly without
even the flimsy excuse of going from a custom product to a standard
product, I suspect they will make a great market opportunity for a
competitor. 

  It's one thing to say "we want to offer a more standard product,
bigger, slower, less reliable, and more profitable, but standard," and
another to say "we're dropping enhancements for what we sold you last
year and converting to something with a cheaper license, even though
it's non-standard."

  My humble opinion is that hackers and techies make up a fairly small
portion of their market, and business people are not going to see a cost
advantage to a "message passing" kernel. A version of X which is not
three years out of date they could see, but the competition is selling
V.3 and V.4, and people are doing useful work on them.

| I have always hated SCO's 3.2 implementation.  OSF/1, on the other hand,
| looks like a dream to me.  A darn good dream.

  What I said about techies. My personal machine is still on Xenix,
shortly to go to V.4. My play system has had a number of things on it,
including a beta BSD and a beta MACH. For getting work done I want
stability and conformance. With V.4 I get it, and symbolic links, too.

| This, coupled with hardware based on the MIPS R3000 and R4000 chips could
| make a serious dent in the Sun installed base.  If these firms back it up
| with world-class support, they have a heck of a shot at being successful
| here.  

  With world class support you can make money selling almost anything. I
wonder if someone will ever try doing that in the UNIX marketplace.

|        The hardware level part (R3000 chipsets) are already fantastic --
| look at the performance available from a MIPS Magnum, for example -- on a
| price-performance level it's a darn nice system.

  Performance I like, but the price is the problem. A 486-33 with 8MB
and 320MB disk, 1024x768x256 display, V.4 with X and compilers, will
still come in under $5k. Maybe this new ACE environment will get a
faster CPU down to this price level, but I don't see it. And new Intel
chips are going to keep getting faster.

|                                                   Add to that OSF/1's 
| functionality and a great support organization (which is dedicated to a 
| great product rather than seeing how much money they can make) and you've 
| got the potential to do some real damage to the Sun empire.

  If you think OSF doesn't have an axe to grind, I'm not going to debate
it here, but I disagree.

| NONE of this is official -- and none of it is from the horse's mouth.  Take
| it with whatever size grains of salt you wish.

  Rumors of a Mach kernel from SCO are everywhere, and they are
officially in the ACE project. I don't think they'll drop V.3 yet
though. I guess I'll agree with your basic thoughts and disagree with
the timetable and details.
-- 
bill davidsen	(davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen)
        "Most of the VAX instructions are in microcode,
         but halt and no-op are in hardware for efficiency"