From: Greg Douglas <gdoug...@reputable.com>
Subject: "Linux is better than any proprietary Unix"
Date: 1999/08/11
Message-ID: <37B1CC7A.D45DDF4F@reputable.com>#1/1
X-Deja-AN: 511550894
Content-Transfer-Encoding: 7bit
X-Accept-Language: en
Content-Type: text/plain; charset=us-ascii
X-Trace: 11 Aug 1999 13:19:12 -0600, 206.168.216.12
MIME-Version: 1.0
Newsgroups: comp.sys.sgi.misc
http://www.eet.com/story/OEG19990810S0019
The emergence of Linux as an industry standard, and the
fact that Linux is better than any proprietary version
of Unix, led SGI to reassess its position in the Windows
NT market, Vrolyk said.
-- John R. Vrolyk, senior vice president of SGI's computer
systems business unit.
Still selling those Origins, John?
This company is doomed.
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Greg Douglas : Linux is only free
Reputable Systems : if your time is worthless.
http://www.reputable.com : - MrNutty.
From: dfev...@bbcr.uwaterloo.ca (David Evans)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/12
Message-ID: <7ov1f8$9la$1@watserv3.uwaterloo.ca>#1/1
X-Deja-AN: 511937568
References: <37B1CC7A.D45DDF4F@reputable.com>
Organization: University of Waterloo
Newsgroups: comp.sys.sgi.misc
Not only that but:
Going forward, SGI will shift its focus to the Linux operating
systems for Intel-based platforms, while maintaining its investment
in and support for MIPS-based systems into 2002, Vrolyk said.
This makes it sound like some time in 2002 support for MIPA IRIX will cease,
meaning that one gets about three years of software support for a shiny new
Origin 2000.
--
David Evans (NeXTMail/MIME OK) dfev...@bbcr.uwaterloo.ca
Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/
University of Waterloo "Default is the value selected by the composer
Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: "Beau Vrolyk" <bvro...@sgi.com>
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/13
Message-ID: <7p2acs$1mq8@fido.engr.sgi.com>#1/1
X-Deja-AN: 512447378
References: <37B1CC7A.D45DDF4F@reputable.com> <7ov1f8$9la$1@watserv3.uwaterloo.ca>
Organization: Silicon Graphics Inc., Mountain View, CA
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Newsgroups: comp.sys.sgi.misc
David,
We have announced that we will be introducing new MIPS based system running
Irix through 2002. Those are the ones that we've formally committed to.
There will be many more after than, but we don't normally announce products
over three years in advance. Those systems will be supported for many years
after their introduction in 2002 and beyond, at least 15 years under the
terms of some contracts. Hopefully, this will clear this up.
Beau
John R "Beau" Vrolyk
SGI
David Evans wrote in message <7ov1f8$9l...@watserv3.uwaterloo.ca>...
> Not only that but:
>
> Going forward, SGI will shift its focus to the Linux operating
> systems for Intel-based platforms, while maintaining its investment
> in and support for MIPS-based systems into 2002, Vrolyk said.
>
>This makes it sound like some time in 2002 support for MIPA IRIX will
cease,
>meaning that one gets about three years of software support for a shiny new
>Origin 2000.
>
>--
>David Evans (NeXTMail/MIME OK)
dfev...@bbcr.uwaterloo.ca
>Computer/Synth Junkie
http://bbcr.uwaterloo.ca/~dfevans/
>University of Waterloo "Default is the value selected by the
composer
>Ontario, Canada overridden by your command." - Roland TR-707
Manual
From: m...@mash.engr.sgi.com (John R. Mashey)
Subject: Re: "Linux is better than any proprietary Unix" [NOT; SGI Chief Scientist]
Date: 1999/08/14
Message-ID: <7p2cb4$kj9$1@murrow.corp.sgi.com>
X-Deja-AN: 512453638
References: <37B1CC7A.D45DDF4F@reputable.com> <7ov1f8$9la$1@watserv3.uwaterloo.ca>
<slrn7r6is8.7gi.alastair@calliope.demon.co.uk> <7ovej7$i77$1@watserv3.uwaterloo.ca>
<37B46F4C.9EB77E5B@broomstick.com>
Organization: Silicon Graphics, Inc.
Newsgroups: comp.sys.sgi.misc
[I'm a Chief Scientist at SGI, started doing UNIX stuff in 1973,
was one of the architects of MIPS, and managed the UNIX port that
turned into both RISC/os & IRIX on MIPS ... so I'm one of the godfathers
of this stuff ... and nevertheless, I am helping us get into Linux on
Intel... & I'll explain why & how].
1) I sit right next to Vrolyk, and in plenty of meetings with him,
and the EETimes quote sounds pretty weird/misquoted to me.
Thsi happens all teh time, sometimes because we don't communicate so well,
and sometimes because press doesn't transmit what is said so well.
2) Rather than trying to answer all these comments piecemeal, why don't
I just tell you what we're doing, in more detail, i.e., from
an engineering view. Some fo this confusion coems from the fact that
textual press releases don't use the roadmap charts I have sitting on my
machines...
3) We announced, in the Spring, at least 2 more MIPS CPUs beyond the ones
on the earlier roadmap. Specifically, the latter of these appears
towards the end of 2002, meaning sold throughout 2003. There is a design
for one more beyond that, but it's not yet committed. I sit in on
design reviews of these things.
4) MIPS-based machines run MIPS-IRIX, and we have a whole stack of
enhancements to IRIX that people are working on. I expect to still be
selling new MIPS-based machines in 2004/2005 - in fact, there's a major,
brand-new system design coming out next year (can't say more) that uses
MIPS-IRIX, and from past experience, what all this means is that we'll still
be enhancing IRIX in 2003, and supporting/maintaining it for years thereafter.
There is an extremely clever [well, I helped a little with it, so I may'
be biased) architecture that will let us cost-effectively continue
MIPS-based systems for a long time, in parallel with some IA-64 machines
that otherwise use the same hardware.
5) During the last few years, we had a whole lot of customers tell us that
while they loved IRIX, sooner or later they were going to go NT;
we had even more ISVs tell us this. It turns out that not all of those
who said that actually did, and it also turns out that a lot of
people keep telling us they'd really love to have Linux systms from us ...
if certain IRIX features were there.
We also discovered there was a strong bifurcation in market between people
who demand that it be NT, and people who demand that it be anything but NT...
and that some of the people who were telling us NT really meant:
lower-cost platform with lots of software, and we don't need every IRIX bell
and whistle, and actually Linux is getting close.
6) Linux *is* better than IRIX in some ways: it's cheaper, and it's getting
terrific application momentum. IRIX is a whole lot better in many other ways,
and *nobody* here is under the illusion that Linux, as it stands right now,
is a direct replacement for IRIX machines, especially for people who use the
cool stuff that IRIX does.
7) We think, by appropriately working within the Linux community, that we
can take important IRIX "genes" and offer them to the community, or
make them addons, and that "over time", we can help Linux be a useful
choice for a higher percentage of people who use IRIX. When I say,
"over time", I mean: there are some things that people do with Irix systems
that could just as easily move to Linux today. There are some other
things where it will be 2005 or later before any Linux-based machine
could be a reasonable replacement.
8) I talk to customers a lot. I show them engineering roadmaps, and it's
a lot easier to do with slides than words, but it makes it 100% clear
that we are:
a) Doing MIPS & IRIX for a *lonnnng time*, and enhancing both.
b) We are firing up LInux work, and providing intellectual
property and work to help Linux, *over time* become suitable for
our customers (at lower cost, and with more applications).
c) It does our customers no good whatsoever if we sell the greatest
OS ever, but it doesn't run the applications they want. We *know*
how to do scalability, high-performance, graphics, 64/32-bit, etc,
and we have a bunch of good people working, soem in IRIX, some in
Linux, and some in IRIX, with the long-term expectation that
those new features will be able to be moved over in a few years.
9) Anyway, my roadmaps show a long period of overlap & coexistence between
IRIX & Linux. There will be IRIX machines still running a *decade* from now.
We aren't going to try to make custoerms do flash cuts or crazed transitions:
we're not nuts; we are going to try to bring some of the great things from
IRIX into the Linux environment, and hopefully getting lower costs and
long-term broader software availability.
Does that help? It doesn't cover everything, and there are still several
things to decide, but what I've described is true to the best of my
knowledge (and it happens that I do the keeping of the top-level roadmaps,
and present them to customers quite often).
[Note: this is generated on an IRIX O2; at home I have an Indy (IRIX),
and a 320 (NT); I will be getting another disk for the 320 so I can dual-boot
it with Linux, but of course, I will personally have at least one IRIX
machine until it dies [if only for old showcase files, and masses of
past history about MIPS, going back to 1985].
--
-john mashey DISCLAIMER: <generic disclaimer: I speak for me only...>
EMAIL: m...@sgi.com DDD: 650-933-3090 FAX: 650-933-4392
USPS: SGI 40U-005,
2011 N. Shoreline Blvd, Mountain View, CA 94043-1389
From: jer...@netcom.com (Jeremy Allison)
Subject: Re: "Linux is better than any proprietary Unix" [NOT; SGI Chief Scientist]
Date: 1999/08/14
Message-ID: <7p2gjl$cmq@dfw-ixnews11.ix.netcom.com>#1/1
X-Deja-AN: 512474558
References: <37B1CC7A.D45DDF4F@reputable.com> <7ov1f8$9la$1@watserv3.uwaterloo.ca>
<slrn7r6is8.7gi.alastair@calliope.demon.co.uk> <7ovej7$i77$1@watserv3.uwaterloo.ca>
<37B46F4C.9EB77E5B@broomstick.com> <7p2cb4$kj9$1@murrow.corp.sgi.com>
Organization: Netcom
X-NETCOM-Date: Fri Aug 13 8:26:13 PM CDT 1999
Newsgroups: comp.sys.sgi.misc
NNTP-Posting-User: jeremy
m...@mash.engr.sgi.com (John R. Mashey) writes:
>[I'm a Chief Scientist at SGI, started doing UNIX stuff in 1973,
>was one of the architects of MIPS, and managed the UNIX port that
>turned into both RISC/os & IRIX on MIPS ... so I'm one of the godfathers
>of this stuff ... and nevertheless, I am helping us get into Linux on
>Intel... & I'll explain why & how].
...The rest of John's post deleted....
Thanks, John for talking about the IRIX plans. I just came from
LinuxWorld Expo where many people were making the same mistake
of thinking that support for Linux means no support for IRIX.
From a completely engineering perspective (ie. no I'm *NOT*
speaking for SGI, just making a personal observation)
what I see here day to day is work being done on Linux in
*addition* to the work being done on IRIX.
I'm employed by SGI to develop Samba (yes, one of those 'evil' Linux
applications :-). Samba is developed first and formost on IRIX.
We test Samba on IRIX benchmark Samba on IRIX, develop new Samba
features on IRIX. When Samba is working correctly on IRIX I then
port it to Linux (RedHat 6.0 actually) in the same way I do to
Solaris, AIX and HPUX. I then additionally benchmark Samba on one
of the new 1400L boxes (that's the new Linux server).
The main samba.org Web server is Apache on IRIX running on an O200 !
As I said, this is just a personal perspective from a developer
in SGI, but all *large* (ie. multi-terrabyte) Samba deployments
that I know about are IRIX based.
Regards,
Jeremy Allison,
(just speaking for myself this time, not SGI or the Samba Team).
From: dfev...@bbcr.uwaterloo.ca (David Evans)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/14
Message-ID: <7p2khr$72v$1@watserv3.uwaterloo.ca>#1/1
X-Deja-AN: 512497750
References: <37B1CC7A.D45DDF4F@reputable.com> <7ov1f8$9la$1@watserv3.uwaterloo.ca>
<7p2acs$1mq8@fido.engr.sgi.com>
Organization: University of Waterloo
Newsgroups: comp.sys.sgi.misc
In article <7p2acs$1...@fido.engr.sgi.com>,
Beau Vrolyk <bvro...@sgi.com> wrote:
>
>We have announced that we will be introducing new MIPS based system running
>Irix through 2002. Those are the ones that we've formally committed to.
>There will be many more after than, but we don't normally announce products
>over three years in advance.
I just read John Mashey's response saying essentially the same thing--good
to hear. It's a shame that the article botched the interpretation as badly
as it did.
--
David Evans (NeXTMail/MIME OK) dfev...@bbcr.uwaterloo.ca
Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/
University of Waterloo "Default is the value selected by the composer
Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: m...@mash.engr.sgi.com (John R. Mashey)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/14
Message-ID: <7p2tv1$pbf$3@murrow.corp.sgi.com>#1/1
X-Deja-AN: 512524591
References: <37B1CC7A.D45DDF4F@reputable.com> <7ov1f8$9la$1@watserv3.uwaterloo.ca>
<7p2acs$1mq8@fido.engr.sgi.com> <7p2khr$72v$1@watserv3.uwaterloo.ca>
Organization: Silicon Graphics, Inc.
Newsgroups: comp.sys.sgi.misc
In article <7p2khr$72...@watserv3.uwaterloo.ca>, dfev...@bbcr.uwaterloo.ca (David Evans)
writes:
|> I just read John Mashey's response saying essentially the same thing--good
|> to hear. It's a shame that the article botched the interpretation as badly
|> as it did.
If you deal with the press a lot, this is the way it goes...
There are good ones, and bad ones, and they're all under tight deadlines,
and dealing with very complicated, fast-changing businesses; we sometimes
don't communicate clearly enough; sometimes we invest energy in educating
reporters, and about the time they get up speed, they switch to a complete
different column area, and we have to start again. But that's life.
-john mashey DISCLAIMER: <generic disclaimer: I speak for me only...>
EMAIL: m...@sgi.com DDD: 650-933-3090 FAX: 650-933-4392
USPS: SGI 40U-005,
2011 N. Shoreline Blvd, Mountain View, CA 94043-1389
From: ca...@bahnhof.se (Carl Bergfors)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/16
Message-ID: <carlb-1608991601200001@pm4-36.bahnhof.se>#1/1
X-Deja-AN: 513377199
References: <37B1CC7A.D45DDF4F@reputable.com>
X-Complaints-To: news@bahnhof.se
X-Trace: dummy.bahnhof.se 934812223 195.178.166.229 (Mon, 16 Aug 1999 16:03:43 MET DST)
Organization: cats
NNTP-Posting-Date: Mon, 16 Aug 1999 16:03:43 MET DST
Newsgroups: comp.sys.sgi.misc
Is Linux better than any proprietary Unix? It does not matter really if it
is right now. Right now IRIX is probably the best one. But in the future
Linux will be the sole Unix os. It is rapidly growing in a rate that will
not be checked by any Unix or Microsoft system.
Many people here are afraid that IRIX will go away but it will eventually
and be replaced by Linux, a Linux that can do everything that IRIX can and
more. It should be clear that people see Linux as a salvation from
Microsoft. The battle is really between Linux and Microsoft in the long
run. In the short run the battle is between Linux and proprietary Unices
but that battle is really over. Linux won. The good thing is that Linux is
the only os that can defeat Microsoft and it will defeat Microsoft. Linux
is Unix with a vengeance. Without Linux the Unix world would have
succumbed to Microsoft. We should support Linux since it can not be
stopped and since its ultimate achievement will be the overthrowing of
Microsoft
Carl
In article <37B1CC7A.D45DD...@reputable.com>, Greg Douglas
<gdoug...@reputable.com> wrote:
> http://www.eet.com/story/OEG19990810S0019
>
> The emergence of Linux as an industry standard, and the
> fact that Linux is better than any proprietary version
> of Unix, led SGI to reassess its position in the Windows
> NT market, Vrolyk said.
>
> -- John R. Vrolyk, senior vice president of SGI's computer
> systems business unit.
>
> Still selling those Origins, John?
> This company is doomed.
>
> --
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Greg Douglas : Linux is only free
> Reputable Systems : if your time is worthless.
> http://www.reputable.com : - MrNutty.
From: m...@mash.engr.sgi.com (John R. Mashey)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/16
Message-ID: <7p9pr1$6ca$3@murrow.corp.sgi.com>#1/1
X-Deja-AN: 513500896
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
Organization: Silicon Graphics, Inc.
Newsgroups: comp.sys.sgi.misc
It is all to easy to become extreme...
1) IMHO, IRIX is one of the best UNIXes out there, and definitely the
best for certain classes of work, and it will be around a long time.
2) Linux has some complementary virtues, and it is much easier to
get IRIX technical goodies into the Linux universe, than it is to
get them into the NT universe, or to get the applications from either of
those places onto IRIX at any reasonable cost.
Maybe it is hard for me to be extreme since I've been doing UNIX work for
>25 years, and have seen a lot of UNIX versions come and go.
--
-john mashey DISCLAIMER: <generic disclaimer: I speak for me only...>
EMAIL: m...@sgi.com DDD: 650-933-3090 FAX: 650-933-4392
USPS: SGI 40U-005,
2011 N. Shoreline Blvd, Mountain View, CA 94043-1389
From: Arthur Hagen <a...@broomstick.com>
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/16
Message-ID: <37B89CB0.D810BF@broomstick.com>#1/1
X-Deja-AN: 513574809
Content-Transfer-Encoding: 7bit
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com>
X-Accept-Language: en,no
Content-Type: text/plain; charset=us-ascii
X-Trace: news.ntplx.net 934845635 204.213.189.141 (Mon, 16 Aug 1999 19:20:35 EDT)
Organization: NETPLEX Internet Services - http://www.ntplx.net/
MIME-Version: 1.0
Reply-To: a...@broomstick.com
NNTP-Posting-Date: Mon, 16 Aug 1999 19:20:35 EDT
Newsgroups: comp.sys.sgi.misc
"John R. Mashey" wrote:
> 1) IMHO, IRIX is one of the best UNIXes out there, and definitely the
> best for certain classes of work, and it will be around a long time.
>
> 2) Linux has some complementary virtues, and it is much easier to
> get IRIX technical goodies into the Linux universe, than it is to
> get them into the NT universe, or to get the applications from either of
> those places onto IRIX at any reasonable cost.
Out of curiosity, how is it easier to port an IRIX goodie to Linux than
a Linux goodie to IRIX?
Regards,
--
*Art
From: EVILjosh <engin...@noorg.org>
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/16
Message-ID: <37B8B04B.635A1656@noorg.org>#1/1
X-Deja-AN: 513600575
Content-Transfer-Encoding: 7bit
References: <37B1CC7A.D45DDF4F@reputable.com> <carlb-1608991601200001@pm4-36.bahnhof.se>
X-Accept-Language: en
Content-Type: text/plain; charset=us-ascii
Organization: noOrg
X-NETCOM-Date: Mon Aug 16 7:40:22 PM CDT 1999
Mime-Version: 1.0
Newsgroups: comp.sys.sgi.misc
> Is Linux better than any proprietary Unix? It does not matter really if it
> is right now. Right now IRIX is probably the best one. But in the future
> Linux will be the sole Unix os.
Doubt it.
> It is rapidly growing in a rate that will not be checked by any Unix or
> Microsoft system.
> Many people here are afraid that IRIX will go away but it will eventually
> and be replaced by Linux, a Linux that can do everything that IRIX can and
> more.
Well, SGI is talking about integrating more IRIX into the Linux base
right? Still though, IRIX is IRIX.
> It should be clear that people see Linux as a salvation from
> Microsoft.
The way you present this, Linux will be the next Windows. Thanks but
I prefer a choice.
> The battle is really between Linux and Microsoft in the long
> run. In the short run the battle is between Linux and proprietary Unices
> but that battle is really over. Linux won.
Right-o. This freys as vendors start adding their functionality to the
Linux code bases.
> The good thing is that Linux is the only os that can defeat Microsoft
> and it will defeat Microsoft. Linux is Unix with a vengeance.
So is IRIX, SunOS, Solaris, DECunix, AIX, UNICOS etc etc etc.
> Without Linux the Unix world would have succumbed to Microsoft.
> We should support Linux since it can not be
> stopped and since its ultimate achievement will be the overthrowing of
> Microsoft
The way you put it, Linux is replacing what Microsoft is.
Yours,
Josh...
_/_/_/ _/ _/ _/ _/
_/_/ _/ _/ _/ _/
_/_/_/ _/ _/ _/_/
From: nos...@oddhack.engr.sgi.com (Jon Leech)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/17
Message-ID: <7pallo$174ru@fido.engr.sgi.com>#1/1
X-Deja-AN: 513650155
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <37B8B04B.635A1656@noorg.org>
Organization: Silicon Graphics Inc., Mountain View, CA
Newsgroups: comp.sys.sgi.misc
In article <37B8B04B.635A1...@noorg.org>, EVILjosh <engin...@noorg.org> wrote:
>> The battle is really between Linux and Microsoft in the long
>> run. In the short run the battle is between Linux and proprietary Unices
>> but that battle is really over. Linux won.
>
>Right-o. This freys as vendors start adding their functionality to the
>Linux code bases.
GPL protects the Linux kernel to some degree; but perhaps more
significantly, nobody is dominant in the Linux market (possible
exception for Red Hat, who is not a hardware vendor and is highly
committed to open-sourcing everything on their distribution).
With nobody dominant, vendors who "innovate" in proprietary and
incompatible ways are likely to be ignored by commercial ISVs who want
maximum market share - and are certain to be ignored by open source
developers. This provides strong pressure to at least get your
interfaces accepted into the mainstream, if not your implementation.
>> The good thing is that Linux is the only os that can defeat Microsoft
>> and it will defeat Microsoft. Linux is Unix with a vengeance.
>
>So is IRIX, SunOS, Solaris, DECunix, AIX, UNICOS etc etc etc.
But unlike etc etc., Linux is not limited to a single hardware
platform and a single OS vendor.
>> Without Linux the Unix world would have succumbed to Microsoft.
>> We should support Linux since it can not be
>> stopped and since its ultimate achievement will be the overthrowing of
>> Microsoft
>
>The way you put it, Linux is replacing what Microsoft is.
As a widely used OS, sure. As a way to dominate the PC market with
inferior design via threats, strongarming, and buyouts, unlikely.
Jon (*not* speaking for my employer)
From: m...@mash.engr.sgi.com (John R. Mashey)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/17
Message-ID: <7pamjt$hi4$1@murrow.corp.sgi.com>#1/1
X-Deja-AN: 513652559
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
Organization: Silicon Graphics, Inc.
Newsgroups: comp.sys.sgi.misc
In article <37B89CB0.D81...@broomstick.com>, Arthur Hagen <a...@broomstick.com> writes:
|> "John R. Mashey" wrote:
|>
|> > 1) IMHO, IRIX is one of the best UNIXes out there, and definitely the
|> > best for certain classes of work, and it will be around a long time.
|> >
|> > 2) Linux has some complementary virtues, and it is much easier to
|> > get IRIX technical goodies into the Linux universe, than it is to
|> > get them into the NT universe, or to get the applications from either of
|> > those places onto IRIX at any reasonable cost.
|>
|> Out of curiosity, how is it easier to port an IRIX goodie to Linux than
|> a Linux goodie to IRIX?
|>
|> Regards,
|> --
|> *Art
In article <37B89CB0.D81...@broomstick.com>, Arthur Hagen <a...@broomstick.com> writes:
|> "John R. Mashey" wrote:
|>
|> > 1) IMHO, IRIX is one of the best UNIXes out there, and definitely the
|> > best for certain classes of work, and it will be around a long time.
|> >
|> > 2) Linux has some complementary virtues, and it is much easier to
|> > get IRIX technical goodies into the Linux universe, than it is to
|> > get them into the NT universe, or to get the applications from either of
|> > those places onto IRIX at any reasonable cost.
|>
|> Out of curiosity, how is it easier to port an IRIX goodie to Linux than
|> a Linux goodie to IRIX?
|>
|> Regards,
|> --
|> *Art
In article <37B89CB0.D81...@broomstick.com>, Arthur Hagen <a...@broomstick.com> writes:
|> "John R. Mashey" wrote:
|>
|> > 1) IMHO, IRIX is one of the best UNIXes out there, and definitely the
|> > best for certain classes of work, and it will be around a long time.
|> >
|> > 2) Linux has some complementary virtues, and it is much easier to
|> > get IRIX technical goodies into the Linux universe, than it is to
|> > get them into the NT universe, or to get the applications from either of
--------------------------------------------------^^^^^^^^^^^^
|> > those places onto IRIX at any reasonable cost.
|>
|> Out of curiosity, how is it easier to port an IRIX goodie to Linux than
|> a Linux goodie to IRIX?
Your words were symmetric, mine weren't.
1) We can take any source code that is SGI IP and offer it to Linux-land,
or do the equivalent work on Linux, or (in some cases) compile our
source and offer it is a binary addon.
2) There are *applications* on IRIX unavailable on NT or Linux.
3) There are many more *applications* that are available on NT,
that are not runnable on IRIX. There are probably less of such on
Linux at this instant, but there are likely to be a lot more very fast,
if I can believe the NDAs I've gotten from software vendors, and if I
believe what I see when I got into Fry's and look on the shelves.
Most of the applications of concern arte those that are only given out
in binary from.
--
-john mashey EMAIL: m...@sgi.com DDD: 650-933-3090 FAX: 650-933-4392
USPS: SGI 1600 Amphitheatre Pkwy, ms 005, Mountain View, CA 94043-1351
From: m...@mash.engr.sgi.com (John R. Mashey)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/17
Message-ID: <7pasat$jch$1@murrow.corp.sgi.com>#1/1
X-Deja-AN: 513676245
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <37B8B04B.635A1656@noorg.org>
<7pallo$174ru@fido.engr.sgi.com>
Organization: Silicon Graphics, Inc.
Newsgroups: comp.sys.sgi.misc
In article <7pallo$17...@fido.engr.sgi.com>, nos...@oddhack.engr.sgi.com (Jon Leech)
writes:
|> With nobody dominant, vendors who "innovate" in proprietary and
|> incompatible ways are likely to be ignored by commercial ISVs who want
|> maximum market share - and are certain to be ignored by open source
|> developers. This provides strong pressure to at least get your
|> interfaces accepted into the mainstream, if not your implementation.
a) Once upon a time, long ago and far away (in NJ), we all had
copies of the source, and we all diddled it, and it got
a bit chaotic. [At one point in 1974/75, a department of 30 of us
had 15-20 *slightly* different versions of the shell, for starters.
This didn't last of course.)
b) UNIX has *always* acted under Darwinian evolution, that is:
a) There is some kind of UNXI that gets popular.
b) It mutates into all sorts of thigns, and then it gets too
crazy.
c) And then, there is selection & consolidation, back to a).
However, unlike the earlier waves of this, there is one thing
strikingly different about Linux (on Intel, anyway): there is
already a pretty strong binary standard that got there without
quite so many of the convolutions that the rest of us did,
and there is building a strong base of binary applications that
platforms are expected to run. This didn't happen:
- during the RISC wars [since there were different architectures]
- during the early X86 & 68K days [I'm not sure, but didn't]
- during the minicomputer days [everybody had the source, and in any
case, binary ISV software wasn't as strong a factor then]
I am not so naive as to think there won't be issues anyway, but I observe
that there is a *much* stronger incentive for multiple vendors to
follow the binary interfaces that are there, at least.
--
-john mashey EMAIL: m...@sgi.com DDD: 650-933-3090 FAX: 650-933-4392
USPS: SGI 1600 Amphitheatre Pkwy, ms 005, Mountain View, CA 94043-1351
From: maple...@gamers.org
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/18
Message-ID: <7pegi7$ab$1@nnrp1.deja.com>
X-Deja-AN: 514209385
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
<7pamjt$hi4$1@murrow.corp.sgi.com>
X-Http-Proxy: 1.0 proxy.uclan.ac.uk:3128 (Squid/2.0.PATCH2),
1.0 mustard.wwwcache.ja.net:8080 (Squid/2.1.PATCH2), 1.0 x36.deja.com:80
(Squid/1.1.22) for client 193.61.250.34, 193.61.255.5, 194.83.240.23
Organization: Deja.com - Share what you know. Learn what you don't.
X-Article-Creation-Date: Wed Aug 18 14:39:05 1999 GMT
X-MyDeja-Info: XMYDJUIDmapesdhs
Newsgroups: comp.sys.sgi.misc
X-Http-User-Agent: Mozilla/4.05C-SGI [en] (X11; I; IRIX 6.2 IP22)
Btw, wrt to the subject line - such a statement is utter
balderdash. You know it, I know, we all know, so I wish
a few senior execs would just come out and SAY it. It's about
time the myth of Linux being the all-encompasing UNIX god OS
was quashed. It's being hyped far too much.
m...@mash.engr.sgi.com (John R. Mashey) wrote:
> 3) There are many more *applications* that are available on NT,
> that are not runnable on IRIX. There are probably less of such on
> Linux at this instant, but there are likely to be a lot more very
fast,
> if I can believe the NDAs I've gotten from software vendors, and if I
> believe what I see when I got into Fry's and look on the shelves.
> Most of the applications of concern arte those that are only given out
> in binary from.
What people are afraid of is IRIX being dropped in favour of Linux.
I do not want that and I believe many others do not want that either.
For the SGI UNIX machines, I want IRIX, not Linux - the latter
is not secure enough or scalable enough IMO.
Has SGI ditched all its IRIX plans? I seriously hope not, or it
means much of the PR which has been released over the past 2
years has been false.
Commitments were made to merging IRIX with Unicos, which I thought
was a fantastic idea.
Commitments were made to scale IRIX to 9000+ CPUs (or have you
all forgotten the original Cray and ASCI press releases?).
Has this been ditched?
I like the GUI IRIX has. I do not like the messyness of Linux.
Btw, why was the volumizer guy laid off recently?
And has the next-gen IR stuff (Bali?) really been scrapped?
It seems to me and many others that the previous PR output which
stated SGI was going to keep its leading gfx tech was false.
Where is Octane's new gfx? It's way behind schedule now IMO.
If/When I ever get an Octane or Onyx2 of my own, I want IRIX
on it with the latest MIPS CPU(s), not Linux running on Intel.
In fact, if SGI *really* wanted the give customers the best
possible performance, it should actually be supporting the Russian
E2K CPU which is MUCH better than IA64 - all the folks over there
need to get it going is $40M. Team up with IBM, get E2K online
and stomp on IA64. Why not? Russian CPUs have been better than
western designs for years.
If SGI really wants to help its *own* products, instead
of switching to everybody else's, why not do the following
two steps which should have been done *years* ago:
1. Make IRIX free, for ALL versions, not just old ones.
2. Make ALL Alias/Wavefront products free to students and academia.
3. Stop asking stupid prices for old products. Currently, Indys,
Indigo2s, etc. are at least 2X more than typical 2nd-hand
prices asked by ordinary 2nd-hand dealers like ITI, MCE,
Reputable, Mashek, XSInt, WUI, etc. and that's *after* the
recent so-called 90% price cuts in Indigo2 stuff.
4. Scrap the license code requirements for products like CosmoWorlds,
CosmoCode, and especially the Impressario Postcript Renderer. Is
SGI *really* making that much money out of these petty license
requirements??
Why is it IRIX 6.5.4 costs $300 in the US (a quote received
by a friend of mine), but more like $940 where I am in the UK?
(a quote I received from SGI UK) Such differences are ridiculous.
All this hoohaa about Linux being the better choice, when all this
time there are many things SGI could have done to help its own
products along. I was told a long time ago by Steve Proffitt that
many of SGI's dumb pricing and configuration practices would be
stopped, yet they still continue today. *Why* for example does
a 4MB TRAM upgrade for Octane/SE cost $4500?? That's insane!
Why are systems still being sold with disks as small as 4GB
and RAM as small as 128MB? Why are O2s still so expensive compared
to the power offered by the VW320 and VW540? And where on earth
is R7000 for O2? (R5200 just doesn't cut it).
Oh John, we want to believe, we really do, but SGI has said some
astoundingly shocking things in recent days when it comes to
reassuring customers old and new.
For over a decade, SGI has maintained order-of-magnitude improvements
in its high-end gfx technology. That era appears to be at an end
for no readily apparent logical reason. Is SGI not even going to
bother fighting against Wildcat4500 and PixelFusion? The more I
talk to people, the more it seems many SGI gfx experts have left for
other companies.
SGI has made commitments and statements in past press releases
about the future of IRIX, MIPS, Cray, etc. IMO it should be legally
bound to stick to those commitments.
Ian.
SGI Network Admin, University of Central Lancashire, Preston, England,
PR1 2HE.
mapes...@yahoo.com | Tel: (+44 -0) 1772 893297, Fax: (+44 -0) 1772
892913
"There is no magic, only stuff." - Nakor, "The King's Buccaneer" (R.E.
Feist)
Doom Help Service (DHS): http://doomgate.gamers.org/dhs/
SGI/Future Technology/N64: http://www.futuretech.vuurwerk.nl/
BSc Dissertation (Doom): http://doomgate.gamers.org/dhs/diss/
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
From: Penio Penev <pe...@venezia.rockefeller.edu>
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/19
Message-ID: <pYHu3.45$Z4.31876@rockyd.rockefeller.edu>#1/1
X-Deja-AN: 514428568
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
<7pamjt$hi4$1@murrow.corp.sgi.com> <7pegi7$ab$1@nnrp1.deja.com>
X-Trace: rockyd.rockefeller.edu 935021461 129.85.41.100 (Wed, 18 Aug 1999 20:11:01 EDT)
Organization: The Rockefeller University
User-Agent: tin/pre-1.4-980105 (UNIX) (IRIX/6.2 (IP22))
NNTP-Posting-Date: Wed, 18 Aug 1999 20:11:01 EDT
Newsgroups: comp.sys.sgi.misc
On Wed, 18 Aug 1999 14:39:05 GMT maple...@gamers.org wrote:
| Has SGI ditched all its IRIX plans? I seriously hope not, or it
| means much of the PR which has been released over the past 2
| years has been false.
| Commitments were made to merging IRIX with Unicos, which I thought
| was a fantastic idea.
| Commitments were made to scale IRIX to 9000+ CPUs (or have you
| all forgotten the original Cray and ASCI press releases?).
| Has this been ditched?
Good questions. But where are the answers?
| I like the GUI IRIX has. I do not like the messyness of Linux.
Very true. But I'd say that inst on LInux is of a higher priority.
| Where is Octane's new gfx? It's way behind schedule now IMO.
Yeah, wasn't this supposed to be announced at SIGGRAPH?
| In fact, if SGI *really* wanted the give customers the best
| possible performance, it should actually be supporting the Russian
| E2K CPU which is MUCH better than IA64 - all the folks over there
| need to get it going is $40M.
With all the intellectual property? I highly doubt it. BTW, is the
E2K sampling? What feature size do they use? Do they use copper?
SOI? Which fabs do they produce it in?
| Team up with IBM, get E2K online
| and stomp on IA64. Why not?
H1 and H2 were better that IA64 too, remember? Why were _they_
dropped?
| Russian CPUs have been better than
| western designs for years.
Why, then, did SGI pay a fine recently for selling computers to the
Russians (and why were they buying it, if they have better ones)? And
why, then, was /Serge moaing and groaning recently that when he was in
Russia, he could not use the latest-and-greatest machines to do
computational chemistry because of the embargo -- didn't he have
access to those better CPUs?
| Why is it IRIX 6.5.4 costs $300 in the US (a quote received
| by a friend of mine), but more like $940 where I am in the UK?
| (a quote I received from SGI UK) Such differences are ridiculous.
This questions you need to address to Her Majesty's Tax Office. SGI
has repeatedly stated that they make _exactly_ the same profit on
every piece of service and/or equipment, regardless of geographic
region.
| Why are systems still being sold with disks as small as 4GB
| and RAM as small as 128MB? Why are O2s still so expensive compared
| to the power offered by the VW320 and VW540?
Because they run IRIX, of course -- you get more value, you pay more
money.
| And where on earth
| is R7000 for O2?
Isn't that a question for QED?
| For over a decade, SGI has maintained order-of-magnitude improvements
| in its high-end gfx technology. That era appears to be at an end
| for no readily apparent logical reason. Is SGI not even going to
| bother fighting against Wildcat4500 and PixelFusion?
But that's the whole story with NVIDIA, right?
--
Penio Penev <Pe...@pisa.Rockefeller.edu> 1-212-327-7423
From: scott@nyetspam_stonebug.net (Scott Elyard)
Subject: If the Russians built it, would they come?
Date: 1999/08/19
Message-ID: <scott-1908991144260001@204.182.133.13>#1/1
X-Deja-AN: 514795867
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
<7pamjt$hi4$1@murrow.corp.sgi.com> <7pegi7$ab$1@nnrp1.deja.com>
<pYHu3.45$Z4.31876@rockyd.rockefeller.edu>
Organization: Jeff, God of Biscuits
Newsgroups: comp.sys.sgi.misc
In article <pYHu3.45$Z4.31...@rockyd.rockefeller.edu>, Penio Penev
<pe...@venezia.rockefeller.edu> wrote:
> | Russian CPUs have been better than
> | western designs for years.
>
> Why, then, did SGI pay a fine recently for selling computers to the
> Russians (and why were they buying it, if they have better ones)?
Just because you have a great CPU design, doesn't mean you've got the fab
facilities to realise it.
> And
> why, then, was /Serge moaing and groaning recently that when he was in
> Russia, he could not use the latest-and-greatest machines to do
> computational chemistry because of the embargo -- didn't he have
> access to those better CPUs?
Not if you can't produce them. And then there's the software itself. I
imagine that's not strictly trivial to develop from scratch.
Russian (and Soviet) high-end engineering has always been impressive,
especially when it comes down to getting the most out of limited
resources. It has always made US examples, by comparison, seem downright
wasteful and inefficient.
SGI teaming up with the Russians. Now that does bring a Grinchy grin to
my face.
--
Scott Elyard ~~~ooOOoo~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
| Am I wrong? Big Deal. The discussion is all that matters.|
| There is no noise like human talk. IRIX, BeOS, & MacOS.|
^~~~~~~~~~~~~~~~~~sc...@nyetspam.stonebug.net~~~~~~~~~~~~~~'
From: Penio Penev <pe...@venezia.rockefeller.edu>
Subject: Re: If the Russians built it, would they come?
Date: 1999/08/20
Message-ID: <ailv3.47$Z4.33115@rockyd.rockefeller.edu>#1/1
X-Deja-AN: 515300340
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
<7pamjt$hi4$1@murrow.corp.sgi.com> <7pegi7$ab$1@nnrp1.deja.com>
<pYHu3.45$Z4.31876@rockyd.rockefeller.edu> <scott-1908991144260001@204.182.133.13>
X-Trace: rockyd.rockefeller.edu 935190790 129.85.41.100 (Fri, 20 Aug 1999 19:13:10 EDT)
Organization: The Rockefeller University
User-Agent: tin/pre-1.4-980105 (UNIX) (IRIX/6.2 (IP22))
NNTP-Posting-Date: Fri, 20 Aug 1999 19:13:10 EDT
Newsgroups: comp.sys.sgi.misc
On Thu, 19 Aug 1999 11:43:34 -0700 Scott Elyard <scott@nyetspam_stonebug.net> wrote:
| In article <pYHu3.45$Z4.31...@rockyd.rockefeller.edu>, Penio Penev
| <pe...@venezia.rockefeller.edu> wrote:
|> | Russian CPUs have been better than
|> | western designs for years.
|>
|> Why, then, did SGI pay a fine recently for selling computers to the
|> Russians (and why were they buying it, if they have better ones)?
| Just because you have a great CPU design, doesn't mean you've got the fab
| facilities to realise it.
Which makes my point -- comprating vaporware (E2K) to sampling silicon
(IA-64) is futile.
--
Penio Penev <Pe...@pisa.Rockefeller.edu> 1-212-327-7423
From: Andrew Maizels <and...@one.net.au>
Subject: Re: If the Russians built it, would they come?
Date: 1999/08/22
Message-ID: <37BF4475.12CA5AB7@one.net.au>#1/1
X-Deja-AN: 515663709
Content-Transfer-Encoding: 7bit
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
<7pamjt$hi4$1@murrow.corp.sgi.com> <7pegi7$ab$1@nnrp1.deja.com>
<pYHu3.45$Z4.31876@rockyd.rockefeller.edu> <scott-1908991144260001@204.182.133.13>
<ailv3.47$Z4.33115@rockyd.rockefeller.edu>
X-Original-NNTP-Posting-Host: 192.168.11.8
X-Accept-Language: en
Content-Type: text/plain; charset=us-ascii
Organization: PixyMisa Mining and Manufacturing
MIME-Version: 1.0
Newsgroups: comp.sys.sgi.misc
Penio Penev wrote:
> Which makes my point -- comprating vaporware (E2K) to sampling silicon
> (IA-64) is futile.
Merced isn't sampling yet. It has taped out, and they expect first
silicon back from the fab in a couple of weeks. Which may or may not
work well enough for engineering samples.
Andrew.
--
Biting weevils by moonlight,
\\\\\ Eating bugs by daylight,
\\\\\\\__. Never running from a food fight,
____\\\\\\\'/_______She is the one named Sailor Hedgehog!
From: Arthur Hagen <a...@broomstick.com>
Subject: Re: If the Russians built it, would they come?
Date: 1999/08/21
Message-ID: <37BF665C.83341C17@broomstick.com>#1/1
X-Deja-AN: 515683436
Content-Transfer-Encoding: 7bit
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
<7pamjt$hi4$1@murrow.corp.sgi.com> <7pegi7$ab$1@nnrp1.deja.com>
<pYHu3.45$Z4.31876@rockyd.rockefeller.edu> <scott-1908991144260001@204.182.133.13>
<ailv3.47$Z4.33115@rockyd.rockefeller.edu> <37BF4475.12CA5AB7@one.net.au>
X-Accept-Language: en,no
Content-Type: text/plain; charset=us-ascii
X-Trace: news.ntplx.net 935290479 204.213.188.108 (Sat, 21 Aug 1999 22:54:39 EDT)
Organization: NETPLEX Internet Services - http://www.ntplx.net/
MIME-Version: 1.0
Reply-To: a...@broomstick.com
NNTP-Posting-Date: Sat, 21 Aug 1999 22:54:39 EDT
Newsgroups: comp.sys.sgi.misc
Andrew Maizels wrote:
> Merced isn't sampling yet. It has taped out, and they expect first
> silicon back from the fab in a couple of weeks. Which may or may not
> work well enough for engineering samples.
Note that Intel usually spends a year from the prototype release until
the production release. It's interesting that HP, who helped Intel
develop Merced, recommends waiting for McKinley instead of going for
Merced. (McKinley is planned to be released during the second half of
2001.) And the MIPS 64-bit CPU's have been around for quite some
time...
Regards,
--
*Art
From: richa...@gno-sbamm.dgii.com (Richard Masoner)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/23
Message-ID: <37c19653.0@samba>#1/1
X-Deja-AN: 516265053
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
Content-Type: text/plain; charset=us-ascii
X-Complaints-To: abuse@visi.com
X-Trace: ptah.visi.com 935433814 199.86.0.206 (Mon, 23 Aug 1999 13:43:34 CDT)
Mime-Version: 1.0
NNTP-Posting-Date: Mon, 23 Aug 1999 13:43:34 CDT
Newsgroups: comp.sys.sgi.misc
Art wrote:
>> 2) ...it is much easier to
>> get IRIX technical goodies into the Linux universe, than it is to
>> get them into the NT universe....
> Out of curiosity, how is it easier to port an IRIX goodie to Linux than
> a Linux goodie to IRIX?
Read Mashey's statement again, Art.
Richard Masoner
From: Arthur Hagen <a...@broomstick.com>
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/23
Message-ID: <37C1D8B5.41513908@broomstick.com>#1/1
X-Deja-AN: 516367849
Content-Transfer-Encoding: 7bit
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com> <37c19653.0@samba>
X-Accept-Language: en,no
Content-Type: text/plain; charset=us-ascii
X-Trace: news.ntplx.net 935450833 204.213.189.191 (Mon, 23 Aug 1999 19:27:13 EDT)
Organization: NETPLEX Internet Services - http://www.ntplx.net/
MIME-Version: 1.0
Reply-To: a...@broomstick.com
NNTP-Posting-Date: Mon, 23 Aug 1999 19:27:13 EDT
Newsgroups: comp.sys.sgi.misc
Richard Masoner wrote:
>
> Art wrote:
>
> >> 2) ...it is much easier to
> >> get IRIX technical goodies into the Linux universe, than it is to
> >> get them into the NT universe....
>
> > Out of curiosity, how is it easier to port an IRIX goodie to Linux than
> > a Linux goodie to IRIX?
>
> Read Mashey's statement again, Art.
Stop misquoting, mr. Masoner. You're making a fool out of yourself,
since everyone easily can go back and look at the original text.
The full text I replied to (and which I quoted and you mangled) was:
> 2) Linux has some complementary virtues, and it is much easier to
> get IRIX technical goodies into the Linux universe, than it is to
> get them into the NT universe, or to get the applications from
> either of those places onto IRIX at any reasonable cost.
Excluding NT from the above statement, I read it as saying it's easier
to port goodies from IRIX to Linux than from Linux to IRIX at any
reasonable cost. Thus my question.
--
*Art
From: m...@mash.engr.sgi.com (John R. Mashey)
Subject: Re: If the Russians built it, would they come?
Date: 1999/08/24
Message-ID: <7psocn$fn4$4@murrow.corp.sgi.com>#1/1
X-Deja-AN: 516382002
References: <37B1CC7A.D45DDF4F@reputable.com>
<carlb-1608991601200001@pm4-36.bahnhof.se> <7p9btl$aqq$1@ffx2nh5.news.uu.net>
<7p9pr1$6ca$3@murrow.corp.sgi.com> <37B89CB0.D810BF@broomstick.com>
<7pamjt$hi4$1@murrow.corp.sgi.com> <7pegi7$ab$1@nnrp1.deja.com>
<pYHu3.45$Z4.31876@rockyd.rockefeller.edu> <scott-1908991144260001@204.182.133.13>
<ailv3.47$Z4.33115@rockyd.rockefeller.edu> <37BF4475.12CA5AB7@one.net.au>
<37BF665C.83341C17@broomstick.com>
Organization: Silicon Graphics, Inc.
Newsgroups: comp.sys.sgi.misc
In article <37BF665C.83341...@broomstick.com>, Arthur Hagen <a...@broomstick.com> writes:
|> Note that Intel usually spends a year from the prototype release until
|> the production release. It's interesting that HP, who helped Intel
|> develop Merced, recommends waiting for McKinley instead of going for
|> Merced. (McKinley is planned to be released during the second half of
|> 2001.) And the MIPS 64-bit CPU's have been around for quite some
|> time...
1) A standard interval is 12-18 montsh from tapeout of a brand-new micro
until you see production systems.
2) Both HP and SGI are doing what any sane systems company does in teh
middle of a transition: offer a long overlap period when both CPUs
are available, so that customers can convert when they feel like it,
which varies tremendously. (Also, both HP and SGI have struggled to
explain this overlap, and it never works very well in linear words,
it's a lot better on graphical chrts where you can see roadmaps.)
3) HP saying *its* customers may wait for McKinley. Note that the
early IA-64s will first appeal to technical-compute customers, who:
a) Often buy early things.
b) Like floating-point
c) Often have a small amount of portable code to move.
They will not appeal so much, at the beginning, to enterprise data managmeent
customers, who:
a) Seldom buy early things.
b) Usually don't care much about floating-point [Wall Stree "rocket science"
types mostly act like technical-compute customers].
c) Needs tons of programs
Both HP and SGI have both kinds of customers, but a much bigger percentage
of SGI's customers are of the "innovator/early-adoptor/early-majority"
technical flavor than are HP's, which means that Merceds are a better
match for SGI's customer base. Compilers (at least SGI's) are already
well-matched to the IA_64 floating-point architecture, whereas every
compiler person I've talked to (not just at SGI) believes that the
integer-side features of IA_64 are fine, but will take a while longer
to get up the learning curve. (i.e., this is the predicated-logic feature,
good for helping gnarly/branchy OS, networking, and DBMS code.)
4) SO, anyway, both of us are overlapping, SGI just has a better match with
customers, and perhaps, with compilers, to take advantage early of IA-64.
--
-john mashey EMAIL: m...@sgi.com DDD: 650-933-3090 FAX: 650-933-4392
USPS: SGI 1600 Amphitheatre Pkwy, ms 005, Mountain View, CA 94043-1351
From: m...@mash.engr.sgi.com (John R. Mashey)
Subject: Re: "Linux is better than any proprietary Unix"
Date: 1999/08/24
Message-ID: <7pt3q9$k76$1@murrow.corp.sgi.com>#1/1
X-Deja-AN: 516438096
References: <37B1CC7A.D45DDF4F@reputable.com> <carlb-1608991601200001@pm4-36.bahnhof.se>
<7p9btl$aqq$1@ffx2nh5.news.uu.net> <7p9pr1$6ca$3@murrow.corp.sgi.com>
<37B89CB0.D810BF@broomstick.com> <37c19653.0@samba> <37C1D8B5.41513908@broomstick.com>
Organization: Silicon Graphics, Inc.
Newsgroups: comp.sys.sgi.misc
In article <37C1D8B5.41513...@broomstick.com>, Arthur Hagen <a...@broomstick.com> writes:
|> The full text I replied to (and which I quoted and you mangled) was:
|>
|> > 2) Linux has some complementary virtues, and it is much easier to
|> > get IRIX technical goodies into the Linux universe, than it is to
|> > get them into the NT universe, or to get the applications from
|> > either of those places onto IRIX at any reasonable cost.
|>
|> Excluding NT from the above statement, I read it as saying it's easier
|> to port goodies from IRIX to Linux than from Linux to IRIX at any
|> reasonable cost. Thus my question.
And I posted shortly thereafter explaining that assymetry of my
statement got lost in the translation. This has nothing to do with
porting, as I'd expect that:
1) Most Linux user source codes port relatively easily to IRIX [and many have]
2) Many applications in the industry can be compiled either place.
3) Some applications (like that use OpenGL seriously, or any of the
IRIX APIs not found on Linux) do not currently port to Linux usefully.
(People are busy adding some of the more popular APIs that IRIX shares
with other commerical UNIX systems.) See, for general info:
http://www.sgi.com/developers/technology/irix.html
and for specific open source work:
http://oss.sgi.com
To try again, more precisely, albeit with more words in [].
it is much easier to get IRIX technical goodies into the Linux universe
[by Linuxizing (?, if that's the right verb) them, and then offering them to
the community as open source, or offering them as binary applications/drivers
that run on/with Linux, either free or for money], that it is to
get them into the NT universe, or to get the applications [usually in
binary form] from either of those places onto IRIX at any reasonable cost.]
--
-john mashey EMAIL: m...@sgi.com DDD: 650-933-3090 FAX: 650-933-4392
USPS: SGI 1600 Amphitheatre Pkwy, ms 005, Mountain View, CA 94043-1351