How did Nvidia infringe 3Dfx texturing patents?
- Thread starter Blito
- Start date
Multitexturing was an inevitability of the industry and it was likely to be be utilised in mutiples of different was (in fat NVIDIA's wasn't entirely the same as 3dfx's in some mode of operation); indeed every manufacturer did it not long after. I'd wager that 3dfx's patent was just sufficiently broad to cater for more than just the individual technical implementations and in this respect, it was ultimately a very good thing for the industry that the trial never came to fruition.
This brings up an interesting questions. How similar is ATIS early adoption of Multexturing? I know ATI has done some wierd things in the past, (3 texture Units per pipe)
I imagine this technique is very similar among all hardware vendors, And I definately agree with Dave here that such lawsuits like these are just damaging to the industry.
Chris
I imagine this technique is very similar among all hardware vendors, And I definately agree with Dave here that such lawsuits like these are just damaging to the industry.
Chris
It also depends on how you discover your insights and protect them. If you can absolutely show you didn't do it using their thinking, patents or approach you can generally protect both your method and approach with a patent - but its tricky stuff.
Digital showed the way when it reversed engineered the BIOS chip thus forming the IBM compatible market. Of course IBM thought with the patent on BIOS it had the market sewn up - just Digital found an independent way right around it that duplicated function.
Patent law is very tricky.
Digital showed the way when it reversed engineered the BIOS chip thus forming the IBM compatible market. Of course IBM thought with the patent on BIOS it had the market sewn up - just Digital found an independent way right around it that duplicated function.
Patent law is very tricky.
All it'd take would be for Texas Instruments to enforce some of their old graphics patents and all the graphics vendors would be in deep doo-doos.
DaveBaumann said:
Can they actually enforce those patents? If its pretty much common knowledge that the techniques in them are being used, then I thought it was the responsiblity of the patent holder to protect their patent. Otherwise, one can just wait till everyone is using their patent and then suddenly ask for licensing fees. Now if TI can show they had no idea that anyone was using their ideas and suddenly found out they could use their patents though. Anyways it all falls under civil law where everything gets fuzzy anyways.
Anyways I doubt TI would do that.
You do NOT need to protect your patent. You can start suing anyone you wish at any time - in fact, that's exactly what is happening today, patent something more or less obvious, then wait until everyone is using it, start suing. That's exactly what happened with the so-called "GIF patent". GIF would never have become a standard image format if the patent would have been enforced from the beginning - simply because there are/were a lot of quite similar compression algorithms around which do not violate this patent and are just as good.Humus said:
(You can lose trademarks if you don't defend them, but not patents. The only way you could lose a patent (apart of the court throws it out because of prior art, triviality or something like that) would be if you actively encouraged everybody to use that technique saying it's freely available.)
Wasn't there talk a little while back about USB key drives? They all have an FAT file system on them and MS recently decided that they owe them for it?
Funny that this thread became an Amiga thread, and while I love the Amiga (used one for 7 years), I want to return to an earlier point.
That was ATI -- a company which introduced pretty interesting innovations (with somewhat flawed implementations), but never provided a way to take advantage of them. And of course a company that made terrible drivers. I wrote my thesis using software OpenGL rendering because the OpenGL drivers were so incredibly buggy. I remember playing Thief on the Rage Pro, where enabling multitexturing caused rendering problems. Ah, those were the days.
ATI actually had multitexturing on the market before 3dfx. The Rage Pro could do some limited combination of 2 textures in one pass, although it was initially available only through CIF, since multitexturing only appeared in DX6, and as far as OpenGL was concerned, ATI initially had no drivers for it, then had really crappy drivers for it, and basically only got multitexturing in when DX6 was already available. Still, that's not half as bad as VQ texture compression, which never appeared outside CIF. So basically ATI was first with both texture compression and multitexturing, but managed to botch them big time.ChrisRay said:
That was ATI -- a company which introduced pretty interesting innovations (with somewhat flawed implementations), but never provided a way to take advantage of them. And of course a company that made terrible drivers. I wrote my thesis using software OpenGL rendering because the OpenGL drivers were so incredibly buggy. I remember playing Thief on the Rage Pro, where enabling multitexturing caused rendering problems. Ah, those were the days.
Voodoo 1 was multitexturing capable - like Voodoo 2 it could cope with up to 3 individual texture processors per rasteriser chip. It was being sold to arcade and vis/sim vendors with these capabilities sometime before Voodoo1 came to the desktop market.
Similar threads
