Slashdot Mirror
OpenGL 1.4 Spec Finalized
Sesse writes: "SGI announced yesterday that the OpenGL 1.4 specification was agreed upon by the ARB. Trying to minimize the gap between D3D8 and OpenGL, the standard adds a lot of functionality already common (being exposed as extensions in many drivers today), but more importantly brings a standard specification for vertex shaders. This should be good news for anybody doing cross-platform eyecandy :-)" This announcement is related to, but broader than, the one mentioned earlier about bringing OpenGL to mobile devices.
SGI is still in charge?
Now I have to upgrade, already
(throws his $500 video card in the garbage)
Je t'aime Stéphanie
Finally, now the T&L unit in my cell phone can be fully used! I was afriad I would have to use DX on my phone for a while there ;)
Unstable Apps: Our Android Apps Don't Suck
why would they have to reverse engineer it?! it is "OPEN".......
MARIJUANA, SHROOMS, X: ONLINE?! - E
Fine, fine, 1.4 is good and all that, but OpenGL 2.0 is where it's REALLY at (as far as game development goes). I'm waiting for the ARB to finally admit that there's two distict uses for GL: CAD and Games. So why not split them off? Well, where would the pressure to extend GL come from if Carmack and Co. weren't shoving Quake N down the IHV's throats? ;-)
Then again, remember MiniGL?
Brr.....
.f00Dave
But SGI sold some of their patents to Microsoft, and I have to wonder if any of them will cause problems for OpenGL 1.4. You know Microsoft isn't about to let OpenGL dominate as the standard for 3-D graphics...
Use 'slashdot stuff' in the subject line in any email you send me if you want to get past the spam filter.
opengl has nothing to do with the kernel ...
:).
....
There is allready a open source clone of opengl called mesa http://mesa3d.sourceforge.net/
So we can assume they will update mesa to suport the new opengl standard (hopefully
Also, once nvidia updates there opengl driver for windows it should be aviabled in there linux drivers too.
Theres no good reason at all to develop a new 3D api, for a start it would takes years to design and implement, and then it would need support from hardware manufactures either in there own implementation or providing specifications for there hardware
Are you fscking insane? There is no "cobbling together" necessary. Just wait for, or better yet help the MesaGL folks to put out the 2.0 spec.
we should work on our own Open Source 3D standard and give it away as Free Software
This is exactly what OpenGL is. An Open specification so that any the same 3D code can run on any hardware/platform/OS.
If it is Open, and succesfull, don't worry, MS will implement it. That's why there is the ICD mechanism on Windows, that all vendors respect.
On Linux, you have the OpenGL ABI that provides the same functionnality. Yes, they would need some more people, but the one they have do a good job.
Now, when you are Open like this, be prepared for competition. The new standard is out, people on the Architecture Review Board have been discussing it for quite a long while. You better have your implementation ready. Or people will go to better support platform.
What you propose is exactly what MS always did : shun away from standards, and try to develop stuff for your platform/OS. The only difference is that you don't have enough market share to do *any* difference.
So, please, cut your crap, follow the standard, and may the best hardware run on the OS with the better/faster support for it.
At the moment, Linux is still in the race (for OpenGL support) and superior in other area. Don't give up.
gah, what am I even bothering to respond...
Anyways, first "Micro$hit" as you so elegantly and maturely call them probably won't be the one incorporating OpenGL into "Winblows X[tra]P[oopy]" (brilliant I must say--share your wit more, please!)--it will be the driver writers from the video card companies.
Secondly, the "kernel hackers at Linux" (wherever that is?!) have nothing to do with this either--ever heard of the Mesa3D project??
Thirdly, reverse engineer? Mesa3D works WITH NVIDIA and others, afaik there's no reverse engineering going on--and not for the base implentation (see sgi).
Fourthly, it's called OpenGL for a reason (hint, pay attention to the open part). In the past this has meant open to users, but it truly is "open" now, check the SGI website for information about the OpenGL sample implementation license. Reverse engineer?? what would you even reverse engineer?? (sorry to repeat myself)
For your last points, feel free to work on your open source 3D standard, if it's useful people will definitely use it. OTOH, DirectX and OpenGL have an immense amount of work put into them already, they might be hard to equal! Oh wait, you said "we" ('we should work') referring to "anyone but yourself"--troll on!
that's ok, he open sourced the doom I 3-d engine, so we can just modify that to use open GL.
um, it doesn't work that way.
First, M$ doesn't give a rip about OpenGL 1.4, and will probably leave nVidia or some other company to do the work of writing to the spec.
Second, Mesa is a free implementation of OpenGL APIs (OpenGL compatible library). I'm sure Mesa's author(s) have already started moving to 1.4.
Third, writing a new Linux library doesn't address the fact that 90%+ of games are written for Windows using DirectX. Creating a new API won't help this unless you do the same thing as OpenGL is doing, which is write cross-platform for Windows, as well as Linux (and maybe other OS's).
Fourth, patents exist on just about anything you can do with computer graphics (CG). OpenGL ARB members share their patents in an effort to make a unified standard and make it affordable for consumers. Microsoft owns some of these patents, mainly through purchased companies, which is why OpenGL Architecture Review Board (ARB) is worried that they could push high licensing fees rather than share patents. It is in Microsoft's best interest to charge money for these patents, especially unreasonable amounts of it, because it makes DirectX the only affordable option and locks you into Microsoft software and x86 hardware.
He didn't say that. Instead, he wrote:
I am now committed to supporting an OpenGL 2.0 renderer for Doom through all the spec evolutions.
"Interesting"? What stupid-ass moderator did that? Do some research first.
If he made such an announcement, it was probably on April 1. The last .plan I've seen is dated June 27 and there's no indication whatsoever of him dumping OpenGL.
Note: I'm aware that the original post was very likely a troll, but I thought that I'd quell the fears of those who took it seriously.
STOP MISUSING APOSTROPHES, YOU MORONS!!!
Bullshit. The last .plan read:
I am now committed to supporting an OpenGL 2.0 renderer for Doom through all
the spec evolutions. If anything, I have been somewhat remiss in not pushing
the issues as hard as I could with all the vendors. Now really is the
critical time to start nailing things down, and the decisions may stay with
us for ten years.
I mean really, this is your worst yet. pathetic... You've always been pretty obvious, but i think this is the worst you've done yet.!!1
How does the M$ patent sabre rattling affect this?
If brevity is the soul of wit, then how does one explain Twitter?
Although OpenGL at times progresses slowly, it's great to see SGI and the ARB actually getting back in the loop and going after D3D. By trying to compete with D3D, it will truly help the future of OpenGL, plus help alternative OS's have 3d software (i.e. Linux, BSD's, and MacOS X).
I brought a poopy basket in case anyone needs to do dirts.
And makes mesa illegal in the US, thus making it impossible to (a) use a linux desktop for gaming and (b) use Mesa to 3d-accelerate eyecandy desktops and (c) use a linux desktop for CAD/Movie production...
I'm not sure if this is an attempt at humour or what, but just in case people take it seriously:
.plans going back to last year sometime, Opengl is still used entirely and will be in future)
John Carmack has never suggested changing to using DirectX, and has always supported ports to other systems.
(According to his
There will probably not be a retail box version of linux Doom3, but a binary port will be releasd as usual, as for the Mac.
Well, there's open and then there's having to pay $$ to qualify for saying your opengl. paying $$ and open doesn't go together.
The most advanced video cards right now are being made for gaming. CAD people will do just fine using a subset of what the gamers use.
This was a PR error by SGI.
The vote has NOT been completed yet.
FUNK!
1. MS does not own OpenGL - it is an open standart
2. MS can not enforce any patents they bought from SGI because when a feature is added to OpenGL all ARB members agree to give their relevant patents under an "ARB Contributor License"(or something like that).
3. MS will have a hard time enforcing any new patents. To quote Neil Trevett from 3Dlabs:
"To affect the creation of a specification, an IP claim must make it impossible to create ANY implementation of the specification that doesn't infringe that IP." You can not patent antialiasing/multitexture/shaders/etc, you can patent only specific alghos that implement that functionality.
I think they mean open because the API is published openly and can be used without fee (with certain restrictions). OpenGL (the name) is a trademark of SGI, so to use it you need to pay a licensing and certification fee.
This reminds me of the whole "free as in beer" and "free as in speech" thing.
Hmmmmm this must be a different .plan file to the one I get when I do finger johnc@idsoftware.com then.
I looked at the opengl.org site, as well as sgi.com and mesa3d.org, but the most recent spec I can find is for OpenGL 1.3
Has the spec been released, or is it only available to ARB members?
Does anyone know a good source for updated development tools (sdks)?
<a href="http://www.opengl.org">www.opengl.org</a> seems a touch or two out of date, linking only to old SGI and MS sdk's (1.2's?). even 1.3's would be a fresh relief.
my opengl coding could use a breath of fresh life.
thanks
myren
open source clone of opengl? opengl is already open !
Never learn by your mistakes, if you do you may never dare to try again
The Doom engine has been GPL'd for, what, 3 years now? Quake was GPL'd about 1.5 years ago, and Quake 2 was GPL'd last December.
Basically, write new engine, wait 6 mos-1 year, GPL old engine.
Want to license an id Software engine for commercial use? Sure, no problem. I think it runs about $1M for the latest and greatest engine (don't know if there are royalties involved as well) or $10k, flat, for an older, GPL'd engine. The key being that you pay $10k and aren't under the GPL, so you don't have to release the modifications back and open up your client to hacks.
3 years from now??? OGL 2.0 is already about 1.5 years behind the times.
Software implementations are fine for proof of concept, but if all Linux had was Mesa for 3D, Microsoft would be all over that by pointing out how their system was superior because it takes advantage of the latest hardware.
As an aside, I find that opengl apps on my Geforce2 tend to run visibly faster in Linux than Windows, on the exact same computer. I would _HATE_ to have no choice for OpenGL compatibility other than Mesa.
File under 'M' for 'Manic ranting'
Not that I'm knocking Mesa, but it _IS_ only a software implementation -- it won't take advantage of hardware like the geforce and radeon.
It will take advantage of earlier hardware though right? That's what the DRI is for? (/me hasn't used mesa with non-3dfx)
Since SGI got the GPL religion, for them to have agreed to the inclusion of the technology in the specification implies that they think the patent is not enforcible, and that their license is still valid.
It would be nice if SGI would state a position on this and clear up the fud, wouldn't it?
-- Terry
What implication does this have on hardware ? Do I now need a new gfx card ? what about opengl 2.0 ? Obviously my card (gf4) supports shaders etc. in hardware ... so the hardware
functionality is there. ???
It is in Microsoft's best interest to charge money for these patents, especially unreasonable amounts of it, because it makes DirectX the only affordable option and locks you into Microsoft software and x86 hardware.
OpenGL would be unaffordable how? nVIDIA already has its own fully-licenced OpenGL drivers for the 3 major OSes (Windows, Mac, and Linux). ATI & Matrox have Windows & Mac covered; the only question is Linux, where neither write drivers. It's not impossible to have MESA implement all non-'patented' OpenGL functions, and the respective hardware makers release the remainder under the (necessary) closed licence.
And more to the point: Windows has a mechanism to allow for other non-DirectX graphics API's. Vid card manufacturers (usually) own full OpenGL licences, and they write complete implementations of OpenGL in their drivers anyway. (Or, to be more specific, they implement the segments of OpenGL that aren't already in their hardware).
Price isn't even an issue, and never was. The cost is shouldered by the vid card makers, and is is hewn down to pennies by the time we pay for it. Neither is x86 hardware-- Or have you forgotten that the primary implementation of WinXP-64bit, which includes DirectX, is Itanium (and while is x86 compatible, it is not x86 or even close to it).
The only real problems that arise is the (expected) moaning that Microsoft is getting money from us whether we buy their software or not, and the future of Mesa or other "Free(dom)" implementations.
And there's nothing from stopping Mesa from implementing everything non-patented, and leaving the patented portions to the hardware makers. Which is still a good deal for ATI or Matrox, as they would only have to write a partial portion of the driver.
For users of nVIDIA and Windows/Mac/Linux, there is and will probably never be a problem; they write their own drivers for all three anyway.
"Free(dom)" software drivers aside, I prefer an excellent, closed-source driver(s) such as nVIDIA's to absolutely no driver at all. It isn't necessarily the HW maker's fault; they have to follow IP laws, and are often kept from releasing source code because of IP laws. If they leave out the 'locked' feature, they lose a competitive advantage, and business to the companies who do. So they choose the best path allowed by law, and provide a non-free driver to a Free OS.
IP law isn't necessarily a bad thing; it's what makes the GPL work. Were it not for IP law, there's nothing from keeping Microsoft from selling our own code back to us.
Information does not want to be free. If it did, we wouldn't have to spend billions in research, either theoretical or applied. People don't give up years of their lives and thousands of dollars to college education because information simply wants to saturate their brains; but because the information requires an active, continuous effort to both spread and simply continue to remain known. Information does everything it can to remain secret. Without our own constant vigilence, all the knowledge and information mankind has collected over the ages would hide iteself again. Skills and facts are forgotten. Books age and crumble. CD's and magnetic media decay.
It takes long, hard work to get information. The whole entropy argument ignores the fact that information is an organized substance, and entropy works against organization, and towards chaos.
While I don't agree on the period of time involved in patents (and espescially copyrights), there has to be a real financial incentive to seek and preserve information. Otherwise, the quest for information and knowledge will be left to rich eccentrics, as was the case centuries ago.
IP law is what made it possible for a person to be a scientist, and earn a living at the same time. It gave them a chance to sell the information they found, and buy their daily bread with the money gained. Without this capability -- to sell the fruits of research and thinking, we would live in a world with very few professional scientists, professional engineers, professional writers (so long to the Lord of the Rings and Dune!) We wouldn't even have flown aircraft yet, let alone flown to the moon.
This does not underscore the greatness of Free Software; it's one of the most altruistic services for all of mankind. But to expect all knowledge to be "Free" is like expecting a farmer to give away his crop.
The world would be nice if everybody shared in this way, but there is a greater human desire to have more if you work more, and that a skilled worker should have more than an unskilled worker. If there isn't an incentive to hard work, study, and the honing of skills, civilization would have never developed.
-- Sometimes you have to turn the lights off in order to see.
Mesa has had hardware renderers since the 3dfx. Back then it was only triangle rendering in hardware, but these days a huge amount of Mesa has hardware paths including T&L. Mesa also forms the core of the DRI and the DRI supports hardware rendering on the Radeon. The following FAQ makes this clear.
I'm not sure where you got the impression that Mesa is software rendering only, but it's untrue.
Seeing is believing.
File under 'M' for 'Manic ranting'
Mesa/DRI does not support hardware acceleration for nvidia cards because nvidia has not been forthcoming with documentation. I agree that seeing is believing, but your conclusion is still incorrect. If you had chosen the correct card and installed the correct software (eg, RedHat 7.3 and a Radeon card) then you would have seen hardware acceleration with Mesa/DRI.