Xref: utzoo comp.misc:6425 comp.sys.ibm.pc:30713
Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!imspw6!bob
From: bob@imspw6.UUCP (Bob Burch)
Newsgroups: comp.misc,comp.sys.ibm.pc
Subject: Re: PC-WEEK Article / OS/2 obituary?
Message-ID: <238@imspw6.UUCP>
Date: 14 Jun 89 15:27:38 GMT
Organization: IMS, Rockville, MD
Lines: 57


From Ted Holden at HTE:
 
.......................
 
From: Gary Barrett, Unisys Corporation, Devon, PA
 
>>
>>
>> From Ted Holden, HTE:
>>
>> Unless I'm reading it wrong somehow or other, it's
>> saying that by 1993, Unix will be doing more than double the business
>> of OS-2 like, basically, hey the game's over; Unix is going to win and
>> OS-2 is going to lose, simple as that.
 
 
> My own feeling is that OS/2 will hit a quick deadend if it does not
> address the following:  1) POSIX, and  2) RISC.
>
> Gates claims to have a POSIX-conformant OS/2 coming down the pike...
 
Let's examine both cases.
 
1)  I am no expert on POSIX, and could easily be off base on this one, but
it seems to me that a POSIX compliant version of OS/2 would have to be called
OS/2-IX or UNIX/2 or some such since it would, in fact, for all intents,
be UNIX, albeit probably a severely ____ed-up version of UNIX.  I mean, a
POSIX manual reads so identically like a UNIX V manual that you assume the
government is merely saying "if you can make your OS function exactly like
UNIX for any and all matters concerning portability (mostly with UNIX systems),
then we might buy it."  And, if it walks like a duck, smells like a duck,
and quacks like a duck.....
 
2)  RISC?  Microsoft has been trying for two years to get a 386 version of
OS/2 together, apparently with little luck.  Kind of like they can't build
a wagon and you're talking about them building Ferrarris and Maseratis.
 
 
What I see really killing OS/2 dead is the following consideration:  that in
the next few years, wherever you go, you'll see mid-sized computers all
running UNIX, database engines (such as the 100 tps Sequent) running UNIX,
workstations like the Suns and Appollos all running UNIX, and scores of 386
and 486 desktop machines, and some poor slob will have to tie all of those
things together.  He's going to think to himself: "Now, do I want those
desktops running OS/2 and me have to deal with two totally different software
worlds forever and with connecting them forever, or would it be simpler
to just run UNIX on everything and use simple uucp connections?"  This
scenario will become the more prevalent as Sequents and machines running
new multi-processor versions of UNIX (the DG or MACH systems for instance)
replace traditional mainframes.
 
I claim that you don't need to be Albert Einstein to figure this one out.
 
Ted Holden,
HTE