[IPOL discuss] Patents and IPOL code

[Nicolas Limare]

Hi,

> There are a growing number of IPOL articles that involve patented
> algorithms, so it is useful to talk about how to handle this
> carefully.  What I say below is my understanding of the situation---I
> am not a lawyer---so please feel free to correct me and contribute
> your knowledge.

Thanks for bringing this issue and your clear explanation.

> # What does it mean if an algorithm is patented?
> 
> For our purposes, there are two concerns:
>    1. Distributing code that implements a patented algorithm
> constitutes patent infringement---even if you completely wrote the
> code yourself.
>    2. Similarly, distributing the program binary of a patented
> algorithm constitutes patent infringement.

I would add that patents are governed by international bodies (WIPO),
international agreements (ACTA), and national laws. What is and is
not legal depends on which set of law applies. IPOL (the editor) is
declared in France and operates from France, so we are governed by
French law apply.

> More specifically, here is what I suggest.
>    1. On the IPOL article, state above the source code links that the
> code is "for research and educational purposes."  (Perhaps we should
> say more here?)

Is the full code "for research and educational purposes", or do these
conditions apply only to the files implementing the patented
algorithm? 

And I need a clarification here: how can we say that the code is "for
research and educational purposes only" and distribute the code under
a free software license who basicaly says "you are free to use, modify
and redistribute this software without restrictions as long as you
keep the license"?

I really don't know is there is a serious conflict or no. GPL licenses
(v2 and v3)[1] have a paragraph about patents, saying that a patent
restriction would cancel any right to redistribute. Some code
implements a patented algorithm and uses the GPL licence
(autopano-sift by Sebastian Nowozin[2]), but I don't know if this is
legally valid. Some patented code is distributed under a BSD licence
with an additional IP grant (webM by Google[3]).

Can someone answer this question: can we distribute the implementation
of a patented algorithm under a free software license without
conflicts between the patent and the license? Can any free software
license be used, or is there a preferred one? 

[1]http://www.gnu.org/licenses/gpl-2.0.html
   http://www.gnu.org/licenses/gpl-3.0.html
[2]http://user.cs.tu-berlin.de/~nowozin/autopano-sift/#download
[3]http://www.webmproject.org/license/

> The source file "dmha.c" implements the algorithm covered by U.S. Patent
> 5629734.  For this reason, the source files "dmha.c", "dmha.h",
> "dmhacli.c" and its binary program "dmha" may only be used for research
> and educational purposes.  Redistribution or commercial use is not
> allowed without the consent of Eastman Kodak Company.
                                 ^^^^^^^^^^^^^^^^^^^^^
Patent rights are bought and sold; I suggest to only refer to "the
consent of the patent owner".

Best,

PS: I think this illustrates how patents may be good for business but
*not* good for research...

-- 
Nicolas LIMARE - CMLA - ENS Cachan    http://www.cmla.ens-cachan.fr/~limare/
IPOL - image processing on line                          http://www.ipol.im/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://tools.ipol.im/mailman/archive/discuss/attachments/20110405/f7dd204a/attachment.pgp>