<strong>IRIX is an amazing OS - with 64MB of ram and a 100MHz MIPS CPU workstation, you can have a killer web server. I have a Sun UltraSPARC 200E (256MB and 200MHz) workstation and it is half as good!
And if you are familiar with file systems, XFS is simply one of the most earth-shatering technologies that has come from SGI (of course, besides OpenGL and the MIPS).</strong><hr></blockquote>
It's really a shame that a company as brilliant as SGI is going down like that - they were really lacking a Jobs. They had the brightest people working for them simply because it was such a cool company (many of then at NVidia now). Too sad, though it only shows how fast a company with good products can go under and it once again proves that Apple knows how to sruvive.
> that. My school has been trying to do some work
> on OpenGL lighting and a couple grad students
> needed access to source algorithims/code for an
> efficient integration of their implementation.
I tought one can download a "sample" OpenGL implementation on developer.sgi.com (or a similiar url)
> Well the answer was no, that is proprietary
> technology, just wrap it even if it kills the
> performance, ... ,maybe we will consider
> assimilating you work into 2.0. Pretty open
> huh? Now read more about "open standards".
Having considered Mesa for your work? That's free (as in GPL)
> There are open standards, and there are
> industry insider standards that are published
> as open but aren't unless you pay a VERY hefty
> sum to sit at the table. The latter is where
> OpenGL currently sits.
Yeah, that's the problem with open "industry" standards - only representatives of the industry can really be there to decide what happens next. And the way to sort people who are serious about it from those who aren't are big amounts of cash.
> If MicroSoft wants to kill OpenGl it can do so
> very effectively but strong-arming Nvidia and
> ATI into DirectX only chip support, Nvidia and
> ATI were already successful in a similar method
> with the death of 3dfx and it's Glide API. The
> rest of the market will follow suit.
If we are talking OpenGL market here (not games) I don't think this is likely to happen. You have companies like Evans&Sutherland relying heavily on OpenGL, making simluators for the US Army and the like (I doubt the US Army would trust it's equipment being shifted towards microsoft technology). You have IBM, HP, Sun, SGI and 3DLabs relying heavily on OpenGL as far as their "pro" products go - microsoft is strong, but they can't dictate the industry _that_ much.
> Sure there may be an incredibly small pocket of
> holdouts, but they will be relegated to ever
> more behind the times technology because there
> won't be enough $$ for good R&D. Anyone
> interested in doing high quality fast 3D work
> will run what is fastest, and if DirectX is it,
> eventually even the biggest objectors will be
> using it lest they "cut their nose off to spite
> their face."
I think (hope?) that what would eventually happen is a OpenGL replacement that is really open - I don't think the whole 3D industry wants their furure in the hands of microsoft's DirectX team.
> How high do you think MS could raise the
> licensing fee for the only economically viable
> 3D API at that point.
high enough for people to come up with an alternative, I guess.
Granted this is all pretty much worst case gloom and doom, but unfortunately it is the case which MS covets most, making it fairly likely.
As for open standards - you can find the <a href="http://oss.sgi.com/projects/ogl-sample/"; target="_blank">Sample OpenGL Implementation</a> source code on sgi.com
> The word sample is the problem, it is not the
> actual implementation for any existing hardware
> set. It is there if you want to write from
> scratch for non-implemented hardware, the
> implemented versions are still proprietary.
> That is not something that is allowable under
> any of the open source licences. Another
> indication that OpenGL is not really "open".
_All_ gfx chip vendors consider their hardware to be "industry/trade secret" I doubt they'll be throwing around OpenGL implementations of them. "Look-ee, here's what our chip can do and this is how it's done." - Not likely to happen. If some people think that this is not what open "standards" should be then they can try and make money off selling gfx chips themselves.
> At about 40 million a pop for a medium fidelity
> flight simulator with motion, they can run
> anything they want.
exactly. and they're not stupid enough to go with M$
> Projects without those kind of resources though
> will be forced to assimilate. SGI sold out so
> they are pretty much toast that hasn't finished
> burning yet;
sad but true
> Sun doesn't NEED OpenGL, Java3D can be wrapped
> around anything they want to wrap it around,
> best value will drive that; IBM/HP will go with
> the largest economical standard as well.
I was thinking of the people using the hardware - the ones who write the apps. If you say "no more OpenGL from now on" then you will most likely scare away all of the customers who used OpenGL on your platform and that means less workstations sold. That is not in the interest of those companies.
> 3DLabs would be the only major potential
> holdout that really would have a stake in the
> standard if M$ tried to kill it. But how long
> could they go it alone if pressure was coming
> from all over?
As long as they make the best 3D graphic cards they don't really need to worry. And they can also - if nothing else helps - licence some of their technology at cheap prices to other companies in exchange to OpenGL backing.
While ATI and NVidia are big and greedy I doubt they'd be stupid enough to say "no" to OpenGL - unless M$ bought both of them. OpenGL is a set standard and M$ can not change that. They tried to replace the internet with MSN and failed and they will fail when they try to replace OpenGL.
OpenGL may not be really that open, but many more companies have a say with it that they would with DirectX, since it would always come down to M$.
Yeah, that's the problem with open "industry" standards - only representatives of the industry can really be there to decide what happens next. And the way to sort people who are serious about it from those who aren't are big amounts of cash.
It's easier to use than coding for PSX 2, especially if the developers only need to port the title from the PC DirectX version.
Also DirectX isn't hard to code with anymore, it used to be at 3.0 but with 8.0 it's actually fun (or so they say). The thing is that stupid microsoft wont port it to OSX or Linux. Oh well.</strong><hr></blockquote>
You would think that if MS was going to write their own 3D library, they would at least improve on what was already out there. But, IMHO, Direct3D has no functionality advantages over OpenGL, yet is much harder to use. Honestly, I have not checked out the 8.0 version though.
COM is only mildly terrible when you're just using COM objects. It gets much worse when you have to write a COM object by hand. And although it's nice if Microsoft provides an IDE to generate COM code, COM should still be clean enough so the programmer is not required to use the IDE, and can fully understand it.
Right now, I am porting an X Server (UNIX C code) to Win32 and cannot use the IDE (have to use nmake and imake). I have needed to turn some structures into COM objects. Doing this by hand (as well as GDI programming in general) has not been fun.
Maybe I am just a perfectionist (see sig). But IMHO, APIs and high level programming languages should be easy to use, and easy to understand without sacrificing much power or efficiency. The Windows APIs, relative to well designed technologies such as Java, Java Beans, OpenGL, or Quartz, are poorly designed, unintuitive black boxes that require alot of brute force trial and error. It's not that it's impossible to write a Windows app, it just takes alot more time (unless you are writing a trivial program), and a lot more testing compared to the well designed platforms out there (PlayStation does not qualify).
Apple needs to do some work to speed up OS X, but, otherwise, they are getting it right with Cocoa, OpenGL, and Quartz. They would never support anything as ugly as DirectX or COM.
Comments
<strong>IRIX is an amazing OS - with 64MB of ram and a 100MHz MIPS CPU workstation, you can have a killer web server. I have a Sun UltraSPARC 200E (256MB and 200MHz) workstation and it is half as good!
And if you are familiar with file systems, XFS is simply one of the most earth-shatering technologies that has come from SGI (of course, besides OpenGL and the MIPS).</strong><hr></blockquote>
It's really a shame that a company as brilliant as SGI is going down like that - they were really lacking a Jobs. They had the brightest people working for them simply because it was such a cool company (many of then at NVidia now). Too sad, though it only shows how fast a company with good products can go under and it once again proves that Apple knows how to sruvive.
> Not really, and the ARB is just a cover for
> that. My school has been trying to do some work
> on OpenGL lighting and a couple grad students
> needed access to source algorithims/code for an
> efficient integration of their implementation.
I tought one can download a "sample" OpenGL implementation on developer.sgi.com (or a similiar url)
> Well the answer was no, that is proprietary
> technology, just wrap it even if it kills the
> performance, ... ,maybe we will consider
> assimilating you work into 2.0. Pretty open
> huh? Now read more about "open standards".
Having considered Mesa for your work? That's free (as in GPL)
> There are open standards, and there are
> industry insider standards that are published
> as open but aren't unless you pay a VERY hefty
> sum to sit at the table. The latter is where
> OpenGL currently sits.
Yeah, that's the problem with open "industry" standards - only representatives of the industry can really be there to decide what happens next. And the way to sort people who are serious about it from those who aren't are big amounts of cash.
> If MicroSoft wants to kill OpenGl it can do so
> very effectively but strong-arming Nvidia and
> ATI into DirectX only chip support, Nvidia and
> ATI were already successful in a similar method
> with the death of 3dfx and it's Glide API. The
> rest of the market will follow suit.
If we are talking OpenGL market here (not games) I don't think this is likely to happen. You have companies like Evans&Sutherland relying heavily on OpenGL, making simluators for the US Army and the like (I doubt the US Army would trust it's equipment being shifted towards microsoft technology). You have IBM, HP, Sun, SGI and 3DLabs relying heavily on OpenGL as far as their "pro" products go - microsoft is strong, but they can't dictate the industry _that_ much.
> Sure there may be an incredibly small pocket of
> holdouts, but they will be relegated to ever
> more behind the times technology because there
> won't be enough $$ for good R&D. Anyone
> interested in doing high quality fast 3D work
> will run what is fastest, and if DirectX is it,
> eventually even the biggest objectors will be
> using it lest they "cut their nose off to spite
> their face."
I think (hope?) that what would eventually happen is a OpenGL replacement that is really open - I don't think the whole 3D industry wants their furure in the hands of microsoft's DirectX team.
> How high do you think MS could raise the
> licensing fee for the only economically viable
> 3D API at that point.
high enough for people to come up with an alternative, I guess.
Granted this is all pretty much worst case gloom and doom, but unfortunately it is the case which MS covets most, making it fairly likely.
I'm currently blowing a cyber-fart in your general direction. :eek:
oooh... cauliflower for dinner last night!
> The word sample is the problem, it is not the
> actual implementation for any existing hardware
> set. It is there if you want to write from
> scratch for non-implemented hardware, the
> implemented versions are still proprietary.
> That is not something that is allowable under
> any of the open source licences. Another
> indication that OpenGL is not really "open".
_All_ gfx chip vendors consider their hardware to be "industry/trade secret" I doubt they'll be throwing around OpenGL implementations of them. "Look-ee, here's what our chip can do and this is how it's done." - Not likely to happen. If some people think that this is not what open "standards" should be then they can try and make money off selling gfx chips themselves.
> At about 40 million a pop for a medium fidelity
> flight simulator with motion, they can run
> anything they want.
exactly. and they're not stupid enough to go with M$
> Projects without those kind of resources though
> will be forced to assimilate. SGI sold out so
> they are pretty much toast that hasn't finished
> burning yet;
sad but true
> Sun doesn't NEED OpenGL, Java3D can be wrapped
> around anything they want to wrap it around,
> best value will drive that; IBM/HP will go with
> the largest economical standard as well.
I was thinking of the people using the hardware - the ones who write the apps. If you say "no more OpenGL from now on" then you will most likely scare away all of the customers who used OpenGL on your platform and that means less workstations sold. That is not in the interest of those companies.
> 3DLabs would be the only major potential
> holdout that really would have a stake in the
> standard if M$ tried to kill it. But how long
> could they go it alone if pressure was coming
> from all over?
As long as they make the best 3D graphic cards they don't really need to worry. And they can also - if nothing else helps - licence some of their technology at cheap prices to other companies in exchange to OpenGL backing.
While ATI and NVidia are big and greedy I doubt they'd be stupid enough to say "no" to OpenGL - unless M$ bought both of them. OpenGL is a set standard and M$ can not change that. They tried to replace the internet with MSN and failed and they will fail when they try to replace OpenGL.
OpenGL may not be really that open, but many more companies have a say with it that they would with DirectX, since it would always come down to M$.
Yeah, that's the problem with open "industry" standards - only representatives of the industry can really be there to decide what happens next. And the way to sort people who are serious about it from those who aren't are big amounts of cash.
[/QB]<hr></blockquote>
Darwin, anyone?
<strong>Originally posted by Willoughby:
> From what I hear, a lot of developers prefer
> the X-Box because of DirectX. Supposedly its
> EASIER to use. But I wouldn't know for sure.
It's easier to use than coding for PSX 2, especially if the developers only need to port the title from the PC DirectX version.
Also DirectX isn't hard to code with anymore, it used to be at 3.0 but with 8.0 it's actually fun (or so they say). The thing is that stupid microsoft wont port it to OSX or Linux. Oh well.</strong><hr></blockquote>
You would think that if MS was going to write their own 3D library, they would at least improve on what was already out there. But, IMHO, Direct3D has no functionality advantages over OpenGL, yet is much harder to use. Honestly, I have not checked out the 8.0 version though.
COM is only mildly terrible when you're just using COM objects. It gets much worse when you have to write a COM object by hand. And although it's nice if Microsoft provides an IDE to generate COM code, COM should still be clean enough so the programmer is not required to use the IDE, and can fully understand it.
Right now, I am porting an X Server (UNIX C code) to Win32 and cannot use the IDE (have to use nmake and imake). I have needed to turn some structures into COM objects. Doing this by hand (as well as GDI programming in general) has not been fun.
Maybe I am just a perfectionist (see sig). But IMHO, APIs and high level programming languages should be easy to use, and easy to understand without sacrificing much power or efficiency. The Windows APIs, relative to well designed technologies such as Java, Java Beans, OpenGL, or Quartz, are poorly designed, unintuitive black boxes that require alot of brute force trial and error. It's not that it's impossible to write a Windows app, it just takes alot more time (unless you are writing a trivial program), and a lot more testing compared to the well designed platforms out there (PlayStation does not qualify).
Apple needs to do some work to speed up OS X, but, otherwise, they are getting it right with Cocoa, OpenGL, and Quartz. They would never support anything as ugly as DirectX or COM.
[ 01-19-2002: Message edited by: Brian J. ]</p>
The day is same day as following event.
<a href="http://www.sgi.com/company_info/investors/events.html"; target="_blank">http://www.sgi.com/company_info/investors/events.html</a>;
SGI will announce second quarter financial results on January 22, 2002, at 2 p.m. PST. Dial-in number is (888) 208-1824 and no passcode is needed.
Hmmmm.....
<strong>Darwin, anyone?</strong><hr></blockquote>
Exactly what does darwin have to do with industry standards?