Hi, I''ve been putting this off for a while but now I''m starting to think it would be best to get it over with. Four months ago I came up with a new, and in many times faster, polygon culling, collision, and sorting algorithm. I''d just like to know how do I get this idea copyrighted so that I can publish it? ---
Brent Gunning | My Site

You don''t copyright techniques, you patent them. A text or graphical display or sequence of symbols can gain copyright protection, a method cannot.

The exact procedure for applying for a patent varies with the country. Google for "patent application" and filter by your home country. Note that not all countries allow method patents (US and most of the EU do).

hth

(My projects and ramblings...)

Okay thanks for clearing that up. So after I get a patent (I live in the US), what do I do? Write a research paper, an article for gamedev.net? What would you do?

---
Brent Gunning | My Site

Write a little paper, publish it somewhere and hope your technique will be adopted by serious developers. Maybe it will be even called ''RapidStunna Culling'' or something, but that will be the largest revenue you get out of it, I''m afraid.
To be liable for a patent, your product needs to be unique, innovative and, most important, have market value. On top of that, a large investment must be made to claim the patent, which depending on the area and duration can easily exceed $10k.
So to get your patent registered, you would need to find an investor that thinks he can sell your product for profit and cough up the patent money.
You may also try to sell it to some third party engine developer, but be prepared the have some doors slammed in your face.

Good luck.

Make profits! Create a product or license the technique to others. There''s nothing special involved just because you have a patent, it''s only there to allow you to make a profit from your own discoveries instead of some competitor just parasiting on others'' ideas.

Note that patent applications cost money - some hundred dollars is my guess (I live in Sweden, where the initial application runs about $500). Then you may need to widen the protection and this requires getting new patents in all areas...
I don''t know the exact cost in the US, but I seem to remember them having lots of problems with processing the applications (like several years'' backlog - during that time is when you see "Pat. pending").

If your idea consists of code or is solved in code, you may need a sample implementation. You may also want to see some advisors to find out if the patent could be applied to other areas (there are companies just going through patent applications to find ways to patent similar things with other applications or contexts).

I''m afraid that''s about as far as I can offer advice - my own idea never made it past talking to patent advisors;-) I recommend you find someone to give you specific advice for your own arena. Good luck!

(My projects and ramblings...)

Though before you even think about approaching a patents lawyer (or approaching the patents offices yourself) you should be 100% certain your "new" method hasn''t already been thought of in the past 30 odd years worth of computer graphics ("prior art") - IMO to be "100% certain" would take around 3-6 months of pure research and reading other peoples papers.

Just because a technique isn''t mentioned on sites like GameDev, it doesn''t mean nobody has already published and possibly patented the idea. There were things published back in the 1970s which are only recently getting widespread use in games and similar "consumer level graphics"...

SIGGRAPH would be the place to publish your paper. The paper approvals board for SIGGRAPH is made up of some of the most knoweledgeable and influencial people in graphics so they''ll a)know if your idea is original, and b)how relevent it is to the world.


Polygon culling, collision & sorting all in one eh. Quite a few things already done in that area. For example Octree/SEAD/BSP etc sorting/indexing to do rough grained culling, collision & sorting has been done to death, and a few patents even exist.

Outcode style culling of batches of polygons were published years ago etc.




Ask yourself WHY you want to patent/copyright something - most techniques aren''t patented etc. If your motivation is purely financial, then its probably pretty pointless to have a dormant patent unless you have the resources to sell & market the technique. You may be better off sitting on the idea, getting employed at a major company/research department and then negotiate a direct, fixed $n000 amount for you to sign your technique to them.


--
Simon O''Connor
Creative Asylum Ltd
www.creative-asylum.com

I think S1CA has the best advice so far. Patent lawyers are expensive, but necessary in some cases. For instance, a patent lawyer might tell you that if you published your idea as suggested in this thread - it would constitute prior art and could jeopardize getting a patent. That would be an expensive lesson. :-( It would also prevent you from being able to protect your method as a trade secret. Common example of a trade secret: The formula for Coca Cola (also the best known trade mark example).

I am an intellectual property attorney but I do not prosecute patents. For this industry I think the cost benefit analysis yields too high a cost for patent protection. An alternative is to copyright it, then license it, and be prepared to litigate it. If it is that good someone will try to steal it. Further, you want to make sure you utilize your state''s Trade Secret protection laws.

Good luck