PDA

View Full Version : SGI patent scare



MikeC
01-16-2002, 11:21 AM
Okay, I guess you've all seen this story - Microsoft buying up a whole shedload of fundamental 3D patents from SGI.

Can any of our resident driver writers (or anybody else, for that matter) comment on whether this is a genuine threat to OpenGL implementors?

John Nagle
01-16-2002, 11:23 AM
Do we have the patent numbers yet?

Adrian
01-16-2002, 11:49 AM
There are some interesting comments on the subject here:
http://slashdot.org/articles/02/01/16/1824256.shtml

Humus
01-16-2002, 12:35 PM
I think people are overreacting a little. Reading this link (http://www.sgi.com/newsroom/press_releases/2001/october/microsoft.html) it looks like more of a matter of M$ buying to rights for something, maybe some tech for DX9, what do I know?
But they say it's a "patent cross-license", not "buying up OpenGL" like lots of people seams to read it.

dorbie
01-16-2002, 03:33 PM
No Humus,

"and the transfer of certain additional intellectual property rights to Microsoft"

This is the press release that tried to snow everyone. It's not about buying up OpenGL, but it is about buying SGI's graphics patents.

Carmacksutra
01-16-2002, 03:39 PM
What about patents on high level shader language stuff?
Does any company have something like this?

dorbie
01-16-2002, 05:24 PM
Yes,

I think SGI has (or had) some shader related patents, I suspect this it one of the things SGI would have held onto, but who knows. Mr Gates et.al. may now own it.

PH
01-16-2002, 11:33 PM
Isn't SGI very dependent on OpenGL ( software systems layered on top of OpenGL, for example ) ? If that's the case ( which I assume ) then they wouldn't sell something that would negatively affect their own products.

opla
01-17-2002, 12:57 AM
I thought these patents are related to game console technology ...

dorbie
01-17-2002, 02:03 AM
Patents are for inventions, in this case horizontal technologies which can be applied in big SGI workstations, PCs or game consoles.

There is no legal difference between a graphics patent used in a workstation and the same graphics patent used in a game console. Heck the hardware is the same in this case too. The XBox graphics system is just a slightly enhanced GeForce3.

Go read that article again, you missed some things I read there. When someone writes 'the bulk of SGI's graphics patents, (or words to that effect), they DON'T mean some piddly little widget developed for the N64. I doubt anything in the N64 gave rise to any new patents at SGI. It might have, but not a significant number of them. This sounds like it's the whole ball of wax. SGI haven't just sold the family silver, they've stripped the lead off the roof and sold it to the guy who burned down their mill.

At least there's no doubt left about SGI's expressed intentions, graphics has been officially declared "non core" at SGI.


[This message has been edited by dorbie (edited 01-17-2002).]

opla
01-17-2002, 04:18 AM
SGI sued ArtX for patent infringement about the design of the GameCube. So, I guess the SGI patents can be game console related. (see http://news.cnet.com/news/0-1003-200-329715.html?tag=rltdnws)

SGI sued also NVIDIA (see http://groups.google.com/groups?q=SGI+ArtX+patent&hl=en&selm=353C1B54.1ED69E15%40metavr.com&rnum=1). Now, NVIDIA has cross licensing agreements with S3 and SGI

dorbie
01-17-2002, 05:35 AM
Oh where to start. It's like you're deliberately trying to confuse the issues or you are very confused.

In context you clearly intended the "related to" phrase to imply that these were SPECIFIC to consoles, that their scope was somehow limited.

That is not the case. Let me repeat, 3D graphics patents are a horizontal technology, i.e. they apply to any 3D graphics product which incorporates them, in any markets, including but not limited to game consoles.

Also, I know that the article mentions patents in the ArtX suit, but there was more to this, there were allegations of misappropriation of trade secrets and other ugly issues in that case which ultimately went nowhere.

The NVIDIA stuff was covered in the article. In anycase that suit and settlement predates their console deal with Microsoft and was about their graphics chips used in PC cards at the time, not consoles. Let me explain, when you have a license agreement you don't have a blank slate to run around and offer the technology to everyone else. You only have the rights granted to you in the license agreement, which might limit all sorts of things, like products, markets or the conditions under which you can sublicense. The patents are still owned by the inventor. In this case the article makes it clear that NVIDIA couldn't just turn around and sublicense SGI's property to Microsoft. Microsoft had to get it's own license from SGI. Infact they didn't just license though, it looks like they completely transferred ownership.


[This message has been edited by dorbie (edited 01-17-2002).]

opla
01-17-2002, 05:54 AM
3D graphics patents are a horizontal technology, i.e. they apply to any 3D graphics product which incorporates them, in any markets, including but not limited to game consoles Yes, as you said "The XBox graphics system is just a slightly enhanced GeForce3"


The NVIDIA stuff was covered in the article. In anycase that suit and settlement predates their console deal with Microsoft and was about their graphics chips used in PC cards at the time, not consoles Correct, it was about the RIVA card.


You only have the rights granted to you in the license agreement. The patents are still owned by the inventor. In this case NVIDIA can't just turn around and sublicense SGI's property to Microsoft. Microsoft has to get it's own license from SGISo, this agrement is/can be for the MS game console. If MS bought these patents because of the XBOX, that's game console related. Or maybe it's for the HomeStation (http://213.40.196.64/content/54/23722.html) but I don't see what are the changes in 3D between the XBOX and the HomeStation.

opla
01-17-2002, 05:58 AM
dorbie, are you working for Silicon Graphics ?

dorbie
01-17-2002, 06:10 AM
No, I used to work for SGI, but this stuff is all in the public domain. This deal happened after I left.

I don't know how many times I have to say this opla. These patents are horizontal technology, it is absolutely clear this is not specific to the XBox. Please be clearer when you phrase this. You say things like 'related to' and 'about'. If you mean specific to then come out and say it. Using these ambiguous terms is just confusing the issue.

If your objective is to say that somehow what Microsoft has obtained here is limited in scope, then you are dead wrong. If you are saying that the XBox was the catalyst for the deal or covered by the patents, that's already in the article and would be self evident anyway.

I'm just trying to avoid confusion here and you seem to be injecting it. If you have a position contrary to this state it, please try and avoid the phrase 'related to' or 'about' stating your position. If you think these patents limited in scope to game consoles then say so, I'll just disagree with you in advance and save myself another post.

Eric
01-17-2002, 06:29 AM
Opla, Dorbie,

It's not that I want to interrupt your little talk ( http://www.opengl.org/discussion_boards/ubb/wink.gif) but can any of you two state clearly what their fears are with regards to this agreement ? I am a little unclear as to what to expect....

Dorbie, you said that this shows that graphics is now "non-core" at SGI. I may be dead wrong but for me, the whole purpose of SGI is graphics... What are they going to do then ????

Sorry for these stupid questions but I'd really like to understand what the big deal is...

Regards.

Eric

opla
01-17-2002, 06:39 AM
Maybe I wasn't clear (English is not my native language).

I wanted to say "the XBox was the catalyst for the deal or covered by the patents" and that this patent deal is not a threat for OpenGL.

Eric, you email doesn't work anymore, I couldn't answer you. Please email me if it's still OK for the pub (I can't thursday 24th)

Carmacksutra
01-17-2002, 07:33 AM
please excuse me this stupid trolling, but i'm getting depressed http://www.opengl.org/discussion_boards/ubb/frown.gif

Is this a future?
patent law == Shariat
IP lawyers == Talibs
IP holders == Mullahs

BTW, Is there a patent on C language?

dorbie
01-17-2002, 11:05 AM
Eric,

I'm making my facetious observation that graphics is "non core" at SGI based on SGI's documented statements about the transfer of "non core" intellectual property to Microsoft in this deal, and nothing else. Feel free to review the evidence for yourself. The alternative explanations seem less generous.

I don't think the sky is falling w.r.t. OpenGL, I don't have any fears I want to expound upon. Most of my posts here have been trying to avoid the facts getting snowed by confused posts & posters.


[This message has been edited by dorbie (edited 01-17-2002).]

Coconut
01-17-2002, 11:18 AM
Originally posted by dorbie:

I'm making my facetious observation that graphics is "non core" at SGI based on SGI's documented statements about the transfer of "non core" intellectual property to Microsoft in this deal, and nothing else.



Dorbie, I don't see the statement mentioned "graphics", how did you know they are patents about "graphics"?

dorbie
01-17-2002, 11:33 AM
Coconut did you read the Register piece?

I've already said I was being facetious, but there's clearly a real unresolved full disclosure issue here. Feel free to rationalize it away in the manner of your choosing.

Funk_dat
01-17-2002, 12:30 PM
There's not really enough information in the article to speculate on underlying business strategy, but I'd guess that it has something to do with Microsoft getting sick of having to pay SGI certain fees everytime they sell a copy of Windows.

Oh..and SGI has kinda been out of graphics for a while now, except for maybe Jon Leech and SGI's ARB duties...SGI has a tiny slice of the server market, basically.

Elixer
01-17-2002, 01:32 PM
You know, I own the patent for when people hit the "reply" button to reply.
I demand code for payment. http://www.opengl.org/discussion_boards/ubb/biggrin.gif

Heh, seriously, I don't think we have enough info yet. It really depends on what they got, and what they didn't, and how long is left on the patent. I just hope it gets cleared up fast.

However, if gfx hardware guys have been told to write the checks to MS instead of SGI for use of OpenGL, then we could be looking at some intresting times.

dorbie
01-17-2002, 06:54 PM
How did you two get from 'we don't know enough' to speculation about Microsoft owning OpenGL?

Microsoft pays royalties on OpenGL, not all the patents mentioned, it's essentially a copyright license (maybe with some IP rolled in I don't know I'm not too familiar with the license details?), AFAIK it's not a per copy license, it's a fairly standard OpenGL license. The actual dollar ammount is peanuts.

We don't have enough information to know the details but we do have SOME information and the signs point to the NVIDIA designed chips in XBox not OpenGL. Let's keep the s/n ratio high by remembering what we do know.

mcraighead
01-17-2002, 10:09 PM
Kids, it's fun to bash IP, but IP makes the tech world go round. Many tech companies, even those that make real products, have much of their real value locked up in IP.

And although I do think patents get misused due to USPTO incompetetence, and that software patents last way too long, they aren't fundamentally all that different than copyrights. One protects the author, the other protects the inventor.

IP is a Very Good Thing (TM). (The pun is unintentional.)

Don't listen to those "information wants to be free" folk. Information is non-sentient and therefore wants nothing of the sort. http://www.opengl.org/discussion_boards/ubb/smile.gif

- Matt

PH
01-17-2002, 11:23 PM
Originally posted by mcraighead:
... they aren't fundamentally all that different than copyrights. One protects the author, the other protects the inventor.


Does Elisha Grey ring a bell http://www.opengl.org/discussion_boards/ubb/smile.gif ? ( the pun is unintentional ). I'm sure there are lots of cases where the wrong inventor gets the rights. Getting a patent means you were faster filing for it than the other smart people out there. The only problem I have with it is that most things are build upon other peoples research ( that may or may not be patented ) and then disallowing others to do the same without paying - that's a legal way to steal.

Of cause I may be wrong http://www.opengl.org/discussion_boards/ubb/smile.gif...

dorbie
01-18-2002, 12:44 AM
Careful Matt,

you might wake up one day to discover that all the patents you've ever filed have been sold to a large predatory monopolist.

dorbie
01-18-2002, 01:36 AM
I just realized that the link to the story which broke the news wasn't posted at the start of this thread. Some readers might be wondering what the heck everyone is talking about. My posts have assumed everyone was on the same page after reading this, here it is incase you need to catch up:
http://www.theregister.co.uk/content/54/23708.html

3DG
01-18-2002, 02:26 AM
Hi

I am a little concerned, do I take it that if MS bought the rights to OpenGL and if so. Now that they have it in the palm of their hand is it possible for them to clench their fist and crush it, and how long could this take. I say this because MS has no real interest OpenGL being in Windows or XBox. And I don’t for one-minute think they have and interest in any other platform that currently supports OpenGL.

3DG

haust
01-18-2002, 02:52 AM
Ok Matt,
so what all this mean for OpenGL future ?? Will MS be able to kick out OpenGL because of copyrights/IPs ownership ??
Honestly i don't care about who owns this and that, i only care about OpenGL independance ;)

Carmacksutra
01-18-2002, 03:29 AM
Originally posted by mcraighead:
IP is a Very Good Thing

I have no slightest doubt about it.

But i think existence of law allowing abusing IP
is just as harmful for innovation, as absolute lack of law protecting IP would be.

BTW, where would be Nvidia now, if 3dfx was too expesive to buy out? (i mean the multitexturing case)


[This message has been edited by Carmacksutra (edited 01-18-2002).]

dorbie
01-18-2002, 03:43 AM
According to the article Microsoft has NOT bought the rights to OpenGL. They have bought some technology patents.

Any talk of Microsoft buying the rights to OpenGL is just baseless speculation. It's probably irrelevant anyway, here's why:

Even if they did, everyone could turn around tomorrow and start a project that was called Free3DG(for example), move the OpenGL 2.0 stuff over to that and get on with their lives. It could even be functionally identical to OpenGL 1.3 and initially use the SGI open source OpenGL implementation and the open source conformance tests (the tests are open source now right?). Owning OpenGL directly would be worthless to Microsoft, especially if they tried to kill it.

OpenGL licensees these days buy the name (trademark), and get stuff like the conformance tests which used to be otherwise unavailable. There is a waiver of some SGI patents in the OpenGL license *I think*, so it's not totally clear cut, especially for new OpenGL licensees but existing licensees should have some rights. Other companies besides SGI occasionally waive I.P. rights when 'donating' spec proposals to the ARB.

In anycase the real concern now is how Microsoft will leverage their patents to influence IHV's like NVIDIA and ATI.

If Microsoft had no hardware I'd be very concerned, because they'd be at least partially immune from the threat of a countersuit, but now they have XBox which is extremely high volume, if they demand royalties or dictate terms it might backfire now because everyone is stepping on eachother's patents.

As for NVIDIA they can't be pursued by Microsoft on the stuff they have already licensed from SGI, but the irony is that because NVIDIA built the XBox, it makes them more vulnerable to Microsoft's unwelcome influence than other IHVs. They have no leverage to counter a threat over any unlicensed SGI patents sold to Microsoft that NVIDIA might use in future products.

They can't countersue Microsoft, they designed the XBox for them!

I wonder what Matt will think of the virtues of patents if Microsoft start dictating terms behind closed doors.


[This message has been edited by dorbie (edited 01-18-2002).]

Carmacksutra
01-18-2002, 04:24 AM
Originally posted by dorbie:
Even if they did, everyone could turn around tomorrow and start a project that was called Free3DG(for example), move the OpenGL 2.0 stuff over to that and get on with their lives.

Actually, all i'm afraid is this is what might be impossible.

It seems you can patent some primary feature like "alpha blending" algorithm: http://www.theregister.co.uk/content/archive/22963.html

PH
01-18-2002, 04:45 AM
The patent can be found here,
http://patft.uspto.gov/netacgi/nph-Parse...RS=PN/5,379,129 (http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=/netahtml/srchnum.htm&r=1&f=G&l=50&s1='5,379,129'.WKU.&OS=PN/5,379,129&RS=PN/5,379,129)

Wow, that's what I call an "invention". Check out that complex, non-obvious equation http://www.opengl.org/discussion_boards/ubb/smile.gif.

dorbie
01-18-2002, 05:16 AM
The fact that you use OpenGL or don't does not protect you from someone's patents. If Microsoft own the right I.P. they can sue you whether or not what you use is called OpenGL. SGI sued NVIDIA over their chip designs and they were an OpenGL licensee sitting on the ARB!

Nobody needs to run off and start a new OpenGL because Microsoft don't own it as far as we know. In anycase they can't recant some of the things SGI has already done with this.

The issue is not whether they own OpenGL.
They own some interesting graphics technology. The next question is how will they use it? We don't know, but the Register is probably correct, it needs some oversight, but what's new? There should have been some oversight when they scuppered the MCD release.


[This message has been edited by dorbie (edited 01-18-2002).]

3DG
01-18-2002, 06:59 AM
Dorbie ok, some good points. I was speculating on the whole owning of OpenGL thing. But MS would surely like to own/destroy OpenGL then again I am looking at this from the whole windows and game programming angle and not the whole multi-platform OpenGL angle. What kind of revenue stream does SGI gain from OpenGL. Would they have cause to sell OpenGL i.e. would it be profitable. Then again if I were CEO of SGI I would want to keep OpenGL just for its brand name alone and the perception of quality it emits.

If MS was the sole owner of OpenGL I do think they could crush it. With the Nvidia example Nvidia must be making $50-70 (maybe I am being a little naive) on each Xbox. I would presume that Xbox will sell 10 million ish units world wide within its lifetime that is big revenue stream for Nvidia. When it is time to dev Xbox 2 I think MS will be able to ask Nvidia to stop being OpenGL compliant (drivers etc) without to much of problem this single stroke would go along way to crushing OpenGL.

Granted Free3DG (I Like that name, not to sure about the free bit though, that was a joke) would be possible if it were included with every app/game that needed it. Or as a download but it would be a big step back and every average Joe would not bother. That non mass-market penetration could spell the end. It would let DX move further and faster aligning it as the number one Graphics API. It goes without saying I hope this does not happen.

One Last Note – Big thanks to all the guy’s at Nvidia I just got myself a GeForce 3, a truly excellent product, my last card was a 3DFX Voodoo 5 5500 a great card but the Nvidia is a lot better. O and excellent dev resources I am so excited at what is possible with the GeForce 3.


3DG

Elixer
01-18-2002, 11:08 AM
You are all failing to see the real picture. I will give you a hint, it has to do with crop circles, and area 51. I can't say anymore.

BigBrother IS watching!

Lev
01-18-2002, 11:54 AM
Well, I just have to throw my thought into the fire:

in the EXT_fragment_lighting, EXT_texture_perturb_normal and other extensions dealing with per-pixel lighting and bumpmapping, there's this statement:


To the extent that SGI has patent rights that are unavoidably
infringed by all implementations of this extension, SGI will, upon
request, grant a license under such patent rights to the requesting
party subject to reasonable terms and conditions, and without
incremental charge or fee. Such license shall be non-exclusive, and
non-transferable, and shall be limited to implementations of the
extension in combination with any conformance certified
implementation of the OpenGL API. Such license is expressly
contingent upon a grant back of a non-exclusive, royalty-free,
perpetual, worldwide license to SGI and its OpenGL licensees under
the requesting party's patent rights that are unavoidably infringed
by all implementations of this extension or OpenGL.

Now the things SGI sold could include these things, but it all rumours, we'll just have to wait.

-Lev

opla
01-18-2002, 12:32 PM
There's no patents on software in Europe anyway, so it's only a USA and Japon problem (if this patents deal is about OpenGL)

Please allow me a personal message:

Eric, there is really a pb with your email:


When trying to deliver your message, the mail server at mail3.brightview.com encountered
permanent problems with the following address:
For <>, could not connect to destination domain (mm-croy.mottmac.com), destination server denied connection

Reporting-MTA: dns; mail3.brightview.com

Final-Recipient: rfc822;<>
Action: failed
Status: 4.4.1
Diagnostic-Code: smtp; 441X could not connect to destination domain (mm-croy.mottmac.com), destination server denied connection
Last-Attempt-Date: Fri, 18 Jan 2002 21:14:18 -0000

dorbie
01-18-2002, 12:57 PM
I wrote fragment lighting application code using these OpenGL extensions and ran them on a simulator, the SGI OpenGL extensions are not how anyone delivers their fragment lighting today.

This means that SGI would not have been obliged to give anyone a free pass on their related patents. SGI offered this to influence designers to follow in their footsteps. The designers declined preferring to deliver a toolbox of lower level, more flexible but difficult to use extensions.

The SGI extensions were basically a monolithic approach which cloned the vertex lighting API for fragment lights, and let you replace various lighting terms with multitexture units. The bump map was a color vector map with the half bias sign everyone uses now, and the tangent and binormal was supplied at each vertex using the coordinate frame, the implementation did all the required vector transformations. This is already in the public domain in the extension details.

The underlying tangent space implementation ideas are documented in a siggraph 97 paper:

Efficient Bump Mapping Hardware
Mark Peercy, John Airey, Brian Cabral

The underlying ideas do seem to cover some aspects of current fragment lighting implementations. I'll leave you to speculate on what SGI may or may not have claimed in patents they may or may not have filed.

It does raise the question of who violates a patent though. Because of the toolbox approach, all NVIDIA are doing is providing hardware capable of dot products using texture components, and other fundamental mathematical operations. It is actually game developers who go ahead and implement the ideas that might be coherently called fragment lighting, although NVIDIA encourages them to do so through samples and advice. If you wanted to sue someone over this you may have to go after game developers and publishers, not hardware manufacturers who merely support dot product operations per fragment. It depends on the patent claims and the legal implcations of encouraging developers through examples and support.


[This message has been edited by dorbie (edited 01-18-2002).]

Nutty
01-18-2002, 01:31 PM
Software patents suck!

Did you know a programmer, or guy can get sued, because he didn't research whether his algorithm is already patented! Even is he came up with the algorithm by himself!

Just because you come up with a neat idea, shouldn't mean someone else is not allowed to come up with the same great idea himself and use it!

I am all for copyright, but patenting software algo's is wrong IMVHO!

Nutty

dorbie
01-18-2002, 01:32 PM
Opla,

these aren't just software patents. SGI has built hardware for years and filed patents while doing so.

Humus
01-18-2002, 02:48 PM
Originally posted by Nutty:
Software patents suck!

I agree fully on that. I generally don't like patents at all, at least not as they are currently working in most countries, software patents sux most of them all. To get a patent I think at least you should have something radically different, something you can really call an invention. That alpha blending patent earlier in this thread is just plainly ridiculous, just about anyone could have come up with it, it's so obvious and it's really nothing but a simple equation, linear interpolation, whatever ...
I find patents to be counterproductive and cause all kinds or problems. Ok, as long as this world is market driven (or greed driven) we'll need a way to protect inventors from being crushed by companies stealing your inventions, but the way patents are set up today it's just causing more problems than it solves. Most inventions are based on work done earlier by other people, so who we really should call the inventor is often not the same as the one getting the patent. And that you can patent the most simplistic things, it just makes me sick. If people believed in sharing and cooperation more than greed and profit this world would be a much better place to live on. Oh well ...

mcraighead
01-18-2002, 09:54 PM
I can believe in sharing, cooperation, greed, and profit all at the same time. Works out best that way. http://www.opengl.org/discussion_boards/ubb/smile.gif (And I would hate to live in a world without greed and/or profit.)

I am not necessarily convinced that software patents are all bad. I think there are some problems with the patent system, but they are not huge problems, and fundamentally patents are OK.

- Matt

jwatte
01-19-2002, 09:32 AM
> There's no patents on software in Europe
> anyway, so it's only a USA and Japon

That's not true, although a widely believed myth.

Just like a software guy can infringe a patent "even though he thought of it himself," a mechanical, electrical, or chemical engineer can do the same. However, those disciplines have gotten used to actually researching patents before shipping highly visible applications.

The idea behind a patent is that if you patent something which is "stupid" then other people will be encouraged to find a better way. If you find a better way, then there is no harm in the patent. If you do NOT find a better way, then there is actual value in the patent -- and it'll be in the public domain after 17 years, instead of being lost as a trade secret forever.

The problems with software patents today seem to be two-fold:

1) Patents (in general), especially in the US, are granted without any stringent "obvious to a skilled practitioner" check, and the "prior art" check typically only checks previously issued patents and published, peer-reviewed research. Many things are considered too simple for publication, but a patent examiner might still consider it non-obvious. The reason the "obvious" test isn't as stringent as we'd like is, as far as I can tell, that the Patent Office can get sued for making the "wrong" determination (according to the filer).

Seems to me that the law isn't being followed, because the whole validity of the patent system hinges on these fairly high bars actually being maintained. But, hey, what do I know?

2) Even though the government grants a monopoly when a patent is filed, enforcement of the patent is left up to civil action brought by the filer. This, more or less, means that patents are worthless unless you're in the $10M-and-up cash balance league. Thus, individuals who come up with good ideas don't patent them; they just ship software, and thus those ideas don't go into the prior art.

Seems to me that if the government actually grants monopolies through patents, they should also provide enforcement. But, hey, what do I know?

I'm just a programmer.

dorbie
01-19-2002, 10:46 AM
Yea let's build a bloated enforcement beaurocracy to blow our tax dollars chasing people over civil issues.

Companies violate each others patents all the time. Do you really want a third party to step in and tell ATI, NVIDIA, Matrox, SGI (oops I mean Microsoft) et.al that they can't use each others I.P? I can guarantee you that every one of these companies will be violating some patent claims of every other. Your suggestion doesn't help innovation, it screws everyone.

The pretext of patents is to encourage progress through investment in R&D, when they start to do the opposite then they need reform, but the reform should be directed at the original goal, not some other hair brained nonsense.

Note a dollar spent on genuine R&D is not the same as a dollar spent trying to circumvent another's patent.


[This message has been edited by dorbie (edited 01-19-2002).]

opla
01-20-2002, 06:21 AM
these aren't just software patents I know dorbie, I was just talking about OpenGL.

I think everybody (except patent lobbies) agree that software patents sucks. There was a debate in Europe about that (see http://www.theregister.co.uk/content/4/14933.html)

dorbie
01-20-2002, 07:05 AM
OpenGL is essentially a trademark now, it's the implementations (which happen to include hardware) which could fall foul of the SGI/Microsoft patents.

I don't know the exact wording, but I think there's also a qualifier to what you are saying. I've read that European software patents which are innovative and technical in nature are permitted.

opla
01-20-2002, 12:10 PM
I've read that European software patents which are innovative and technical in nature are permitted.
it was out of question to allow business patents like the amazon 1-click patent anyway. there was no question about it: it's bad.
to allow "innovative and technical in nature" software patents, the patent lobbie will have to prove that it's good for innovation.
patents are not the only way to protect innovation. many people (including me) think that copyright and author's right are enough.

dorbie
01-20-2002, 01:45 PM
Without commenting on what the law should be, I think the current status of the law in Europe is such that you can file software patents if they meet certain requirements. IANAEL

V-man
01-20-2002, 07:38 PM
It seems to me that's its difficult to prove that someone has violated a software (or algorithm) patent.

If someone figures out a solution on there own, produces an implemetation that appears to have identical behavior to your own, how can you sue them without solid proof that they violated your patent?

It's not like that someone has stolen your source code or your object code. That person put a lot of effort to produce his version.

V-man

Nutty
01-21-2002, 12:35 AM
Exactly V-man! But it can happen!
It it your responsability to make sure _your_ code doesn't infrigne upon someone elses patent.

But, I dont know of a single person who actually cares about this stuff here in the UK.

Nutty

opla
01-21-2002, 12:56 AM
Remember when Lotus sued Borland for having the same menu tree ?
Even if you win or lose, if you're a small company, you can't spend much money to defend your rights ... http://www.swiss.ai.mit.edu/6805/articles/int-prop/lotus/lotus-borland-appeal-Mar95.html

kehziah
01-21-2002, 01:59 AM
It makes me think of Creative Labs who sued Aureal.

AFAIK Aureal actually won the trial, but went bankrupt because it spent all its money to pay lawyers.

Patents good for innovation?? Aureal's A3D was innovative, and far better than Creative's EAX. Using the crappy patent law, Creative just crushed an innovative competitor to keep its monopolistic situation.

it'll be in the public domain after 17 years
What a joke. Everybody knows that 17 years in computer science is like 100 years in other fields. Remember the state of the art in 1985? I don't, I was too young http://www.opengl.org/discussion_boards/ubb/wink.gif .

Keermalec
01-21-2002, 06:37 AM
SGI insisted at the time these are "non core" technologies, but sources close to the Mountain View are emphatic that these represent the bulk of SGI's 3D intellectual property assets, a view confirmed by documents disclosed to The Register. http://www.theregister.co.uk/content/54/23708.html

I think that about says it all. Nothing is confirmed as yet.

dorbie
01-21-2002, 10:46 AM
V-man, you don't have to 'prove' anything, you just have to convince a jury of 12 people in Delaware (or some other place)that you are right, there's no 'beyond any reasonable doubt' here. Now remember that Bill Gates beat out Thomas Edison as the greatest ever U.S. inventor in a recent poll of the kind of people who will be on that jury. Infact it's marginally worse, because if some local knows anything about computers and computer graphics you will have no chance of being allowed on the jury.

Keemalec, what are you saying? The documents the Register has confirm it, they claim. This is the story everyone's discussing, and for who knows what reason, you are posting a link as if it calls into question the story it presents, it does not. Your selected quote is the most emphatic part of the article, they have documentary CONFIRMATION of the story. Didn't you read this right?

[This message has been edited by dorbie (edited 01-21-2002).]

OldMan
01-22-2002, 12:25 AM
That depends. Microsoft has initiated a lawsuit against a Brazilian company becouse of a suposed violation of micosoft code. And is not going to get anything..becouse American laws are not aplicable here. And since micosoft didn`t payed a process that the Brazilian government has upon it... Well... it can`t get anything here. The law here says that MicroSoft would need to free all their code to show that there is a violation, that is the only way to sue a Braz. company. That is just an example.

Jon Leech (oddhack)
01-22-2002, 10:26 PM
Originally posted by dorbie:
The fact that you use OpenGL or don't does not protect you from someone's patents. If Microsoft own the right I.P. they can sue you whether or not what you use is called OpenGL. SGI sued NVIDIA over their chip designs and they were an OpenGL licensee sitting on the ARB!

As is often the case with legal issues, the situation is complex. OpenGL licensees (a category which includes Microsoft, one of the original licensees, but I digress) agree to a reciprocal licensing agreement for necessary IP to implement OpenGL. Similarly, when the ARB approves a new extension or version of the core specification, ARB members and participants must either raise IP objections within a defined window after voting, or by not doing so, agree to reciprocal licensing of their necessary IP.

However, necessary IP is not all IP. Frex, when 3dfx patented multitexture hardware, this raised a potential problem for ARB_multitexture. 3dfx eventually agreed that there were commercially reasonable implementations of hardware supporting multitexture which did not infringe their patents. This was sufficient for the ARB to adopt the multitexture extension without concern that anyone implementing ARB_multitexture would necessarily be at risk. However, 3dfx continued to pursue NVIDIA because they felt that NVIDIA's implementation did infringe their patents.

Necessary disclaimer: I am not speaking as a lawyer or speaking for my employer, just citing some of the details of the licensing and standards process, and one particular example where the process worked as it was supposed to for OpenGL as a whole.

Jon Leech
SGI

dorbie
01-23-2002, 03:08 AM
Thanks for the clarification, it's marginally better than I would have hoped, I was aware that various contributors frequently waived I.P. rights to extension proposals, I didn't know that it was formalized like this. It seems to make the riders on some extension proposals redundant.

Unfortunately the word necessary makes quite a difference, particularly when it comes to hardware accelerated implementations. Is it necessary for OpenGL to be fast? Is it necessary for example to have an on chip texture cache?

I just want to clarify for anyone that may otherwise get the wrong impression that being an OpenGL licensee doesn't give you rights to this kind of fundamental I.P. and you're potentially screwed without it.

This seems to confirm that the issue is graphics patents and not OpenGL ownership, with an emphasis on implementation details and non ARB approved extensions.

Thanks again for the informed post.


[This message has been edited by dorbie (edited 01-23-2002).]

Jon Leech (oddhack)
01-23-2002, 12:12 PM
Originally posted by dorbie:
Unfortunately the word necessary makes quite a difference, particularly when it comes to hardware accelerated implementations. Is it necessary for OpenGL to be fast? Is it necessary for example to have an on chip texture cache?

Terms like necessary and commercially reasonable often have precise legal meanings that are well removed from how they're used casually by engineers. Working this stuff out within an IP framework defined by a standards body is a large part of why standards bodies exist at all.

But for the most part the disputes that have escalated to public notice have to do with hardware patents, not APIs. The ARB as a whole shouldn't get in the middle of arguments between individual members over the implementation details of their chips; the disputes would be much the same even if completely different vendor-specific APIs were being used by the parties to the dispute. What the ARB needs to do is define APIs that are reasonably possible to implement without tramping on IP that companies want to protect.

Jon Leech
SGI

dorbie
01-23-2002, 01:22 PM
On chip texture cache is a hardware patent, but there are many software aspects to OpenGL implementations, and it's all the more interesting when you consider the application of middlewear code that might sit on OpenGL but not be part of the implementation some of which is now perhaps working its way into OpenGL 2.0.

I don't believe in excluding ourselves from a meanigful discussion just because a lawyer used a word and not an engineer. There is empirical evidence beyond books on contract law of what the word means for OpenGL implementors, the word is still English and interpreted by law. I just need to look at the disputes that have risen to public attention to understand (in my unsophisticated engineer's fashion) about the caveat 'necessary' in this context. You are aware of why I deliberately chose the example I did of course.

I wonder if Microsoft will show up to the next ARB meeting with a lawyer.


[This message has been edited by dorbie (edited 01-23-2002).]