I hope no one from M$FT is reading this
Steve Cordova
Item: Intel is thinking (planning) on moving from x86 architecture
(CISC) to new RISC architecture. M$FT legacy apps would be broken
(Win32 APIs).
Item: Both M$FT and Intel exhibit paranoid behavior as a part of
corporate policy. (See Grove's book and Gates' behavior. It is sound
corporate behavior as much as we Mac Advocates would rather deny this
fact.)
Item: Connectix Inc. releases chip VM for X86 architecture (soon
to include MMX emulation) that runs at approx. 75% of x86 speed on PPC
chip.
Item: Connectix Inc. CTO believes fervently that "Real Men Code in
Assembly"
Synthesis: If Intel is planning on changing to RISC architecture, both
Intel and M$FT have deep enough pockets to buy Connectix, or commission
at sufficiently high price an x86 VM for new chip to retain backward
compatibility.
Conclusion: _Apple Must License VPC_ and figure a way to insure that
Connectix does not write a tightly coded VM for either Intel or M$FT.
Failure to do so allows M$FT to continue its forward and backward
compatibility marketing strategy.
IMHO: Connectix is the ONLY company capable of producing a well coded,
tightly written VM that would allow this.
Comments and E-mail Welcome.
Flames disregarded after first paragraph.
Steve Cordova
scor...@dnvr.uswest.net
scor...@linux1.west.denver.k12.co.us
e-mail will be answered.
Eric Bennett
In article <3351BC...@dnvr.uswest.net>
Steve Cordova <scor...@dnvr.uswest.net> writes:
> IMHO: Connectix is the ONLY company capable of producing a well coded,
> tightly written VM that would allow this.
DEC has already written exactly the kind of emulator MS needs (FX/32
for the Alpha version of NT), so Connectix is clearly not the only
company that could do this sort of thing.
--
Eric Bennett ( er...@pobox.com ; http://www.pobox.com/~ericb )
Sixty-seven percent of the doctors surveyed preferred X to Y. (Jones
couldn't be persauded.)
-John Allen Paulos
Ted Oliver
In article <5itfhj$1q...@r02n01.cac.psu.edu>, er...@pobox.com (Eric
Bennett) wrote:
> In article <3351BC...@dnvr.uswest.net>
> Steve Cordova <scor...@dnvr.uswest.net> writes:
>
> > IMHO: Connectix is the ONLY company capable of producing a well coded,
> > tightly written VM that would allow this.
>
> DEC has already written exactly the kind of emulator MS needs (FX/32
> for the Alpha version of NT), so Connectix is clearly not the only
> company that could do this sort of thing.
>
Slight topic shift - - and I am not disagreeing with your point - - but
aren't these two different things? The Connectix Virtual PC product is
indeed a virtual machine (x86) on which one can allegedly run any x86
operating system. AFAIK, !FX-32 is a binary translation program - it
translates x86 Win-32 programs dynamically into Alpha instructions, and if
I remember, it saves the translated instructions so the translated
programs run faster the second, third, etc. time the Win-32 program is
run.
Steve Cordova
Ted Oliver wrote:
>
Eric Bennett wrote:
>
> > Steve Cordova wrote:
> >
> > > IMHO: Connectix is the ONLY company capable of producing a well coded,
> > > tightly written VM that would allow this.
> >
> > DEC has already written exactly the kind of emulator MS needs (FX/32
> > for the Alpha version of NT), so Connectix is clearly not the only
> > company that could do this sort of thing.
> >
>
> Slight topic shift - - and I am not disagreeing with your point - - but
> aren't these two different things? The Connectix Virtual PC product is
> indeed a virtual machine (x86) on which one can allegedly run any x86
> operating system. AFAIK, !FX-32 is a binary translation program - it
> translates x86 Win-32 programs dynamically into Alpha instructions, and if
> I remember, it saves the translated instructions so the translated
> programs run faster the second, third, etc. time the Win-32 program is
> run.
>
Neither post is IMHO incorrect nor off topic. However, I was trying to
express a
scenario about the near future, and was trying to solicit responses (not
just
trolling) to a situation that I could possibly foresee. Connectix'
proven
track record of providing an acceptable(?) solution to a situation or
opportunity
within the smallest mRAM commercially possible can either stand or fall
by
itself.
Jesper Lai Petersen
Steve Cordova wrote:
>
> Item: Intel is thinking (planning) on moving from x86 architecture
> (CISC) to new RISC architecture. M$FT legacy apps would be broken
> (Win32 APIs).
The processor you refer to is the upcomming Merced processor.
According to a recent Infoworld article,
http://www.infoworld.com/cgi-bin/displayStory.pl?970411.wmerced.htm ,
that processor will maintain compatibility with existing code.
I think you could be quite sure that WinTel are'nt going to screw up
their market lock during this transition.
>
> Item: Connectix Inc. releases chip VM for X86 architecture (soon
> to include MMX emulation) that runs at approx. 75% of x86 speed on PPC
> chip.
I think we should waint and see some actual benchmark tests before
jumping to such conclusions.
> Item: Connectix Inc. CTO believes fervently that "Real Men Code in
> Assembly"
> Synthesis: If Intel is planning on changing to RISC architecture, both
> Intel and M$FT have deep enough pockets to buy Connectix, or commission
> at sufficiently high price an x86 VM for new chip to retain backward
> compatibility.
>
> Conclusion: _Apple Must License VPC_ and figure a way to insure that
> Connectix does not write a tightly coded VM for either Intel or M$FT.
> Failure to do so allows M$FT to continue its forward and backward
> compatibility marketing strategy.
>
> IMHO: Connectix is the ONLY company capable of producing a well coded,
> tightly written VM that would allow this.
Hmmm...Really?
just a little hint:
http://www.macweek.com/mw_1114/nw_lismore.html
-Jesper