creative labs teaches Carmack a thing or two about shadows?

The only way this is a valid patent is if they are claiming undisclosed invention as a way to contest the prior art from Sim Dietrich (he did a presentation on it before Carmack and before the filing date, more info here). Since he says he posted about this method for the first time on a developer board at Creative this makes me very suspicuous.

That they filed for the patent ala, but now they are trying to shake down developers? Fuck that, any self respecting website should nail these fuckers to the wall if this is true. Starting with B3D hopefully :) If Reverend is reading this, could you ask Carmack about this perhaps?

A company which is so closely linked to the gamer community should know better ... this is unacceptable behaviour IMO. They can jump high and low about the law, but fact remains even if their patent is legally valid the fact that the method was independently invented before their patent was even filed and again before it was published shows the patent is not just. I think websites in a position to punish unjust behaviour which sets such a dangerous precident should not be fooled by talk of legality. Just cause its legal doesnt make it right.

Marco

PS. the US patent system is so open to abuse ... allowing undisclosed invention as a way to gain a patent even if someone else discovers the same independently and produces prior art before you file is so open to abuse, there are a few downsides to first-to-publish as far as patents are concerned, but nothing compared to the stupidity of the US system.
 
Simon F said:
SteveHill said:
Ironically, there is a really elegant solution to this clipping/capping problem... but it is also patented.
Click to expand...
Sounds interesting; would you mind elaborating or digging up a link? (I've not had much luck with patent searching in the past.)
Click to expand...
It was used in the Elan T&L chip that was part of Naomi2 arcade system. The basic idea is that as the clipping is actually being done against the front clip plane, the generated edges are used to construct the cap.
Click to expand...
There is another elegant solution for that problem, which isn't patented AFAIK: depth clamp
 
Chalnoth said:
Well, if you ask me, his little writeup that has been posted at nVidia's site for ages would be reason alone for that patent to be thrown out. Too bad it so damned hard to get a patent thrown out....
Click to expand...
What's that got to do with it? The patent was filed before Carmack re-invented it** and that is what counts in such matters.


**to the best of my knowledge but I'd be happy to see evidence to the contrary.

MfA said:
PS. the US patent system is so open to abuse ... allowing undisclosed invention as a way to gain a patent even if someone else discovers the same independently and produces prior art before you file is so open to abuse, there are a few downsides to first-to-publish as far as patents are concerned, but nothing compared to the stupidity of the US system.
Click to expand...
I'm not sure what you are saying here but, IIRC, in this case the patent was filed before publication.

Xmas said:
There is another elegant solution for that problem, which isn't patented AFAIK: depth clamp
Click to expand...
I didn't think that was terribly elegant at all :
(a) You can't use the standard projection techniques
(b) You potentially need a lot more fill rate.
 
Im talking about Sim's presentation which is from march 1999. As I said, the only way the method from that presentation could be covered by the patent is if Creative uses undisclosed invention as a way to establish their rights. That is possible since patent rights in the US are based on first to invent ... a very poor system indeed IMO.
 
I have a question : what if Carmack didn't described his algorithm ? Everyone would have thought that it was just basic shadow volumes, so the owner of the patent couldn't say anything. So basically software patents will keep programmers from describing their techniques from fear of being sued by violating someone else patent. what a source of progress... :rolleyes:
 
Simon F said:
Chalnoth said:
Well, if you ask me, his little writeup that has been posted at nVidia's site for ages would be reason alone for that patent to be thrown out. Too bad it so damned hard to get a patent thrown out....
Click to expand...
What's that got to do with it? The patent was filed before Carmack re-invented it** and that is what counts in such matters.


**to the best of my knowledge but I'd be happy to see evidence to the contrary.
Click to expand...
From what I understand, a patent can be shown to be invalid if their is either prior art, or it can be shown that the idea is obvious. Since Carmack obviously came to the solution independently, I think the second could easily be argued.
 
Are you all sure that this blurb about "shadowing" isn't talking about some kind of audio technique that Creative is using in EAX rather than a visual rendering technique? Seems like id/creative is only adding EAX options to Doom 3.
 
Nope, defintely referring to a patent Creative got a while back. id Software hasn't been known (well really Carmack as far as I know that is) for liking Creative. Doubt id Software would have let Creative see the source code to add EAX in unless they are being forced somehow.
 
MfA said:
PS. the US patent system is so open to abuse ... allowing undisclosed invention as a way to gain a patent even if someone else discovers the same independently and produces prior art before you file is so open to abuse, there are a few downsides to first-to-publish as far as patents are concerned, but nothing compared to the stupidity of the US system.
Click to expand...

This is the big problem with the US patent system. They seem to take the view of allowing almost everything, and then making people go to court to sort it all out at a later date. This discourages the smaller inventors, and heavily favours the big companies with deep pockets that can out-lawyer the little guy. This is exactly what the patent system was supposed to prevent.
 
MfA said:
Im talking about Sim's presentation which is from march 1999.
Click to expand...
OOOH... I hadn't seen that... interesting.
As I said, the only way the method from that presentation could be covered by the patent is if Creative uses undisclosed invention as a way to establish their rights. That is possible since patent rights in the US are based on first to invent ... a very poor system indeed IMO.
Click to expand...
It's more likely that this presentation was never raised as relevant prior art. Theoretically it is the duty of the inventor to submit any new discoveries to the patent office.... but....
 
Chalnoth said:
From what I understand, a patent can be shown to be invalid if their is either prior art, or it can be shown that the idea is obvious. Since Carmack obviously came to the solution independently, I think the second could easily be argued.
Click to expand...
"Obviousness" is usually dealt with in the examination phase whereby the examiner says that, by "simply" putting together X and Y, you get the invention.

Given how long stencil shadows have been in use and how long it was before the Z-fail idea came about I don't think "obvious" is applicable. <shrug>
 
Simon F said:
Xmas said:
There is another elegant solution for that problem, which isn't patented AFAIK: depth clamp
Click to expand...
I didn't think that was terribly elegant at all :
(a) You can't use the standard projection techniques
(b) You potentially need a lot more fill rate.
Click to expand...
Depends on how you define elegance. I'd say it's more related to ease of use than speed. E.g. Python is slow, but terribly elegant, IMO :D
Depth clamping is one of these "it just works" things. But limited to NVidia HW...

(a) Which standard projection techniques do you mean?
(b) True, but not that much more if you use scissoring cleverly.
 
I remember talking to a corporate lawyer and IIRC, they told me that one can't patent math. One could use this to get around patents on software algorithms -- which shouldn't be allowed in the first place. Copyrights on algorithms is entirely different.
 
Nope, math can be patented. See RSA patent, which is essentially a patent of Euler's totient function and discoveries in number theory.
 
RSA is an encryption algorithm, so yeah the algorithm is patented. However, you _could_ likely pull an AMD talk at a different level of abstraction, in this case the math and get away with a copy cat.

Lets say I did something funky in, I don't know, differential geometry, I couldn't patent it, I could patent a process in which it was used, but not the math ALONE.
 
Nope, RSA pretty much boxed off the entire idea of public key encryption. People tried variations, using different mathematical operations, and still got sued.

The US patent system is screwed. I mean, comeon, 1-click, "shopping carts", "online auctions", and pretty much "online anything" got patented. Netscape got sued by Wang because Wang had patented the "Save As" menu option!
 
You must log in or register to reply here.
Back
Top