Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!aurora!labrea!bloom-beacon!oberon!pollux.usc.edu!papa
From: papa@pollux.usc.edu (Marco Papa)
Newsgroups: comp.windows.x
Subject: Re: Mit X conference and who has the PC/RT server ?
Message-ID: <6722@oberon.USC.EDU>
Date: 4 Feb 88 18:06:51 GMT
References: <2007@elrond.CalComp.COM> <6532@oberon.USC.EDU> <1152@polyslo.UUCP>
Sender: nobody@oberon.USC.EDU
Reply-To: papa@pollux.usc.edu (Marco Papa)
Organization: University of Southern California, Los Angeles, CA
Lines: 54
Keywords: X on PC

In article <1152@polyslo.UUCP> jeff@polyslo.UUCP (Jeff Weinstein) writes:
>
>In article <6532@oberon.USC.EDU> papa@pollux.usc.edu (Marco Papa) writes:
>	Actually, the Megapel server started out with about 10 low level 
>	display dependent routines written by the ACSC.  They were  poorly 
>	written, slow and dependend on their own hacked kernel mods that 
>	caused our kernel people to cringe.  We threw most of these routines
>	out in the first week, and as of several months ago they were all 
>	gone.

Sure it was "hacked" code, but it bootstrapped in 2 hours with the Palo
Alto cfb.  Also we had megapels way before Palo Alto had them, and I believe
the work we did definitely cut down on the rest of the development.  Otherwise,
why would IBM want it?

>>                   The IBM MIT group that showed it in Kresge was one of the
>>original beta testers of the Megapel server.  
>
>	The server shown at MIT was being shown by Erik Fortune and
>	Tom Paquin, who along with myself developed the server at
>	IBM Academic Information Systems in Palo Alto.  These were
>	the actual developers, not beta testers.

I did not see neither Tom nor Eric.  A guy that was claiming to be working 
with Berkowitz of the IBM MIT group handled the machines when we where there.

>	The rumor I heard was that when people came to the IBM display
>	at EDUCOM they asked why the server they had just seen at the
>	ACSC was slower than the one on display at the show.

That was after we had stopped working on the low-level server routines.  IBM
Palo Alto was changing the code every day, and we did not have the latest,
faster version.

>>The server runs on an IBM RT PC with IBM/4.3 (Not to be confused with
>>the AIX X-11 version announced for September 1988).  As such it is still an
>>unreleased and unannounced product.  (Though, you won't have to wait long :-)
>                                                 ^^^^^^^^^^^^^^^^^^^^^^^^^
>	Gee, commenting about unannounced IBM products?  Better look over
>	your shoulder Marco, here come the IBM cops.  :-)
>
>	FYI: The Megapel/AED/APA-16/EGA server shipped yesterday.  As far
>	as I know this is the first production quality X11 color server
>	on the market.

FYI I knew that long ago, and note that the date of my posting was BEFORE
the announcement of Feb. 2nd.  And yes, it is the first production quality
X11 color server on the market, and I claim that it probably would not
have been if we (ACSC) hadn't done the work we did on it.  But that is a
personal opinion.

By the way, IBM called it "IBM - ACSC joint development". I did not.

-- Marco
   USC Advanced Computing Support Center