A Lawyer's View on the OpenGL Patent Mess

PDAJames writes "This article has an interesting take on Microsoft's claims on OpenGL technology. An IP lawyer says that Microsoft could make things difficult for OpenGL if they feel like it, basically. "

Comments

The case for OpenGL

[stevenbee]

From the article:

OpenGL owes its current success to an unrestrictive licensing system, which allows developers to write to the API without a licence, and which carries no royalty fees.

I think they were a little remiss in overlooking the technological case for OpenGL; the fact is that many developers prefer it to DirectX, and not for ideological reasons.

Re:The case for OpenGL

[ewhac]

the fact is that many developers prefer it to DirectX, and not for ideological reasons.

Unfortunately, this is not as true as we might wish it to be. Micros~1, as usual, has stacked the deck in its favor.

If you're a graphics card manufacturer, you want to make sure that your drivers will do the right thing. Thus, you need to test the hell out of it. Surprise, surprise, Micros~1 has a little thing called WHQL (Windoze Hardware Quality Labs) which exhaustively tests your hardware and drivers to make certain they behave properly. If you pass, you get a little WHQL sticker to put on your box. Thus, the graphics OEM can save the expense of creating its own verification department.

What do you have to do to earn WHQL certification? Well, that changes over time as Windoze acquires more cruft^H^H^H^H^Hfeatures. But one thing you must do is fully complete their DirectX validation suite. If you fail, you don't get the sticker.

So who does OpenGL validation? Well, Micros~1 will do that for you, too. But it's optional, not required. Moreover, they won't perform OpenGL testing unless and until you've already passed DirectX testing. So, if you're a graphics OEM living on razor-thin margins, you're not going to spend one engineering dollar more than is absolutely necessary to get that WHQL sticker, and to heck with everything else. So OpenGL gets short shrift.

So why did OpenGL get anywhere at all? Two words: John Carmack. Carmack and id Software are the de facto certification authority for OpenGL: "If Quake runs, it works." Trouble is, earlier versions of Quake only used a subset of the full OpenGL API, so card makers only supported exactly that. As Carmack exercised more of the API in new releases of Quake, card vendors slowly got the idea that supporting the complete API was probably a wise move. NVidia got the hint way early, and it didn't hurt that they had a bunch of ex-SGI engineers on staff.

But even so, OpenGL support remains spotty and uneven, because there is no comprehensive certification authority (that wields any political clout) for OpenGL. If your DirectX implementation is broken, Micros~1 will tell you exactly what you messed up. There is, to my knowledge, no such facility in place for testing OpenGL. Thus, OpenGL implementations are broken in different ways across different cards. DirectX is fundamentally broken, but because of WHQL testing, it's broken the same way across all cards. Because of this comparative uniformity across cards, game developers just go straight to DirectX, and maybe will write an OpenGL rendering layer as an afterthought, despite the fact that OpenGL is easier to write for, and can often be seen offering higher performance. id Software was the sole exception to this rule, offering OpenGL support only. It looked like CroTeam, creators of Serious Sam, were going to boost OpenGL's mindshare, but they have since caved in, and their latest Serious Sam release features DirectX support.

Now, Micros~1 has acquired the lever it needs to kill OpenGL. Vertex shaders are the Next Big Thing in hardware-assisted rendering, and they have been under development for some time. It was hacked into DirectX as of DX8 (IIRC), but the OpenGL ARB has been trying to come up with an equivalent solution that is cross-platform and network-transparent. (Hence the perception that OpenGL is "lagging" DirectX.) Even if the ARB makes vertex shaders an optional extension, it will effectively kill OpenGL's already-tenuous popularity as a rendering API, because developers won't be able to rely on vertex shading extensions being installed. Thus, if you're a game developer, and you want vertex shading in your game, then you'll use DirectX and nothing else.

This is Micros~1's idea of, "Competing on the merits."

Schwab

Trying not to bash Microsoft...

[MarvinMouse]

But don't they just make it way to easy lately.

Like, I don't mind other companies that have tons of products and are making tons of money. Plus, they may have a somewhat stranglehold on industries. But I would have to say the only reason I dislike Microsoft is their apparent philosophy of don't produce good products, kill the competition, and use lawyers as much as possible to help both of the above.

If MS produced quality products, I wouldn't care much about their attempts at complete world domination. But, since they don't produce quality products because they don't have to with the monopoly they have. (Remember Bill Gate's quote from some book I read recently which said (approximately), "You don't want them to want your product, you want them to think they cannot survive without your product. Then you win." Or in rough translation, "Don't worry about creating good products, just manuever yourself into a position where they have no choice but to use your products."

Seems about right for MS lately. (Again, I really am not trying to bash Microsoft, just frustrated with what they have been doing.)

Re:Incredible Insight!

[Yohahn]

Actually one of the most interesting things that was pointed out was the question of the legality of selling something that was already agreed to be released in an open manner.

They agreed to release the patent to OpenGL, so can they sell it to somebody who is going to "un-release" it?

If this is what Microsoft and SGI did, it's an interesting problem.

How to Take Over the 3D Industry in a Ten Steps

[MBCook]

Here is Microsoft's plan (as I see it):

1. Get computers to the point where 3D is a possiblity - Done
2. Get computers to the point where 3D is common - Done
3. Notice a competitor/3rd party owns the dominant 3D standard - Done
4. Develop your own standard (Direct3D maybe?) - Done
5. Refine it to the point where it's actually useable - Done
6. Help make many of the important features of modern 3D and get it in competitor/3rd party's standard - Done
7. Point out that you have patents/etc on those parts of the standard and that you will charge large licensing fees on using that standard - In Progress
8. Use fee to strangle the competing standard - To Be Done
9. Now everyone is forced to use your software for 3D if they don't want to pay tons of license fees - To Be Done
10. Watch as competing platforms (let's call them Fruit Computers, and Penguindynamics) die under licensing fees becase you refuse to put your royalty-free API on their platforms - To Be Done
11. Have a good maniacle laugh - To Be Done

See how simple that was?

Re:How to Take Over the 3D Industry in a Ten Steps

[ryants]

# Develop your own standard (Direct3D maybe?) - Done

Except MS didn't develop the original DirectX API. It was developed by a company called RenderMorphics (and it was called something like Reality Lab 3D) and Microsoft just bought it.

This, apparently, is "innovation".

Their claims are probably invalid

[dh003i]

The most likely scenario here is that they bought IP from SGI which SGI had given to the OpenGL project under a public/OSS/FS license. Thus, MS' claims are invalid. You can't put something into the public domain and then take it back. Sorry, that's just not permitted. Once something's in the public domain, its there forever.

MS has pure Hatred towards OpenGL

[ShwAsasin]

It's obvious that Microsoft hates/fears OpenGL. Since the beginning release of 1.1 for Win95, Microsoft has done nothing about releasing the stub/dll source-code, updating the source-code, or even trying to progress the development of OpenGL.

Time and time again they have attempted to copy and improve upon OpenGL, first with Fahrenheit/XSG, then with DirectX. Yet, through all the technology and resources Microsoft puts in, the masses still like OpenGL.

The principals of OpenGL are the same as the day it started with IrixGL. Keeping it simple, functional, and cross-platform. Although Microsoft has gone great strides with DirectX API, they have nowhere near the simplicity of OpenGL. And with the Alternative OS's supporting OpenGL (Mac OSX, Linux, FreeBSD, Solaris, Ps2Linux, etc...) there are more emerging platforms which they cannot touch.

Game/Multimedia Developers are starting to realize that Linux and other platforms are decent for games, and are developing software for them. No, Linux isn't going to take over the world tommorrow because it has OpenGL, but think of this: If a developer gives both the Linux Binary and the Windows Binary, wouldn't you be curious to compare speeds between the two? People would problably spend the extra $0-5 difference for a dual-os game starting the eventual craze. It only takes a few people/companies to start a revolution.

Microsoft is trying to attack every angle of the industry to focussing our attention on their superior product, yet nothing screams superior when their is a true choice and competition in the market.