Newsgroups: comp.sys.next
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!psuvax1!ukma!usenet.ins.cwru.edu!eagle!data.nas.nasa.gov!data!gumby
From: gumby@Cygnus.COM (David V. Wallace)
Subject: NeXT to go with 88K (?)
References: <151480@pyramid.pyramid.com>
	<1991Apr10.215125.28932@neon.Stanford.EDU> <2473@fornax.UUCP>
Date: 12 Apr 91 06:04:09
Organization: Cygnus Support, Palo Alto, California
Sender: news@nas.nasa.gov
In-Reply-To: oneill@fornax.UUCP's message of 11 Apr 91 17:29:06 GMT
Message-ID: <GUMBY.91Apr12060409@Cygnus.COM>
Lines: 41

   Date: 11 Apr 91 17:29:06 GMT
   From: oneill@fornax.UUCP (Richard Oneill)

   In article <151480@pyramid.pyramid.com>,
      tsych@pyrthoth.pyramid.com (terry sych @ pyramid technology corp.) writes:
   >from san jose (ca) mercury news (bits & bytes) 4/10/91:
   >
   >[...]  a next source said company engineers are hard at work
   >on the new next machine, which will be based on motorola's
   >yet unannounced 88110 risc microprocessor.  

   i'm probably opening a can of worms here, but anyone care to comment on
   the ramifications of this decision. not with regard to *what* risc processor
   they have chosen, but just the fact that they are going to be changing
   processor architecture at all.

NeXT has in the past played its cards close to its chest -- this is
the first time I've seen "a next source" quoted in an article.  The
NYT thing, on the other hand, claimed that it was a Moto source who
told them NeXT had chosen the 88K.  Is that because NeXT chose the 88K
or because NeXT, while investigating what chip to use next, talk to
Moto?

Speculating is fun (I do some below) but I wouldn't lay my money on
anything without hearing it from NeXT.

   does this mean that anyone buying a 68040 next today is buying something
   that will very soon be obsolete. in my department we will soon be throwing
   out sun3's because aparrently sun isn't keen on supporting them any more.
   is this going to happen with next and, if so, after how long.

For folks with source I doubt it will be an issue for a while.  After
all, the next compiler is the GNU compiler which can be retargetted to
a variety of machines.

And as far as the 88K goes, it turns out that you could make a tool
which would convert 68K binaries into 88K binaries (as long as they
ran completely in user space).  This is hard to do in general; it's
just due to certain architectural similarities that you can get away
with it in this particular combination.  Who knows about performance
though...