[IPOL discuss] Patents and IPOL code

[Nicolas Limare]

> -systematically, the patented algorithms published in IPOL will be
> sent to their authors as referees.

It will be difficult. How would you contact John F. Hamilton and James
E. Adams who were working for Kodak in 1995? And how can we know if
an algorithm is patented? The fact that we didn't find an patent
doesn't mean that no patent exists.

> In case of touchy patent owners, we will only transmit the code to
> the referees under NDA. IPOL will therefore run online a code
> certified by referees, but which will not be public.
> 
> In other terms IPOL will only publish (actually following its
> definition) "Algorithms", with as a bonus an implementation
> certified by referees, and running on line. The code of this
> implementation will be certified but not public.

And no one can access the source code? I think this makes IPOL much
less interesting if the readers only get a description (even a precise
one) the algorithm.

> Nevertheless, it is not enough to write "for research and
> educational purposes": we must write "for non profit research and
> nonprofit educational purposes", because there is lucrative research
> and there is lucrative education.

Lucrative or non-profit is not relevent. Being non-profit is not an
exception to patent laws. For example, you can't product and give for
free anti-AIDS drugs if they are protected by patents.

> I disagree with Nicolas' proposal to put a BSD or other licenses,
> and invite users to decide by themselves if it is legal or not
> depending where they live. Putting a license at this point would be
> confusing. As a publisher we are supposed to know the law. It is our
> responsibility to avoid any ambiguity about rights.

If we want to know the law for sure, we need a patent attorney to
check if, in France, a local patent or a recognized international
patent is restricting our rights to distribute a source code
implementation for every of our algorithms. Patent restrictions depend
on the local laws, so we can't just suppose that, because a patent was
filed 15 years ago in US, Canada, Sweden or Japan, no one can
implement it in France. Doing this patent investigation is a hard
work, and I think it is not ours.

Just some other examples to show how complex this problem is:
* in 1998, the Fraunhofer Institute sent a letter to MP3 Software
  Developers: "To make, sell and/or distribute products using the
  standard and thus our patents, you need to obtain a license under
  these patents from us." 
  -> http://www.chillingeffects.org/N/464
  -> http://www.mp3-tech.org/patents.html
* the LAME Project, a free software implementation of a MP3 encoder
  under GPL license, states that: "Many people believe that compiling
  this code and distributing an encoder which uses this code would
  violate some patents (in the US, Europe and Japan). However, *only*
  a patent lawyer is qualified to make this determination. The LAME
  project tries to avoid all these legal issues by only releasing
  source code [...]. Source code is considered as speech, which may
  contain descriptions of patented technology. Descriptions of patents
  are in the public domain. [...]
  Note that under German Patent Law, §11(1) a patent doesn't cover
  private acts with non-industrial purposes. Probably interesting for
  developers is that a patent doesn't cover acts with experimental
  purposes, that aim at the object of the patented invention (§11(2))."
  -> http://lame.sourceforge.net/tech-FAQ.txt
* the European Patent Convention says "shall not be regarded as
  inventions [...] schemes, rules and methods for performing mental
  acts, playing games or doing business, and programs for computers"
  -> http://www.epo.org/law-practice/legal-texts/html/epc/2000/e/ar52.html
* the French law excludes "les programmes ou séries d'instructions
  pour le déroulement des opérations d'une machine calculatrice"
  -> http://www.droit.univ-paris5.fr/warusfel/articles/JurInvLog_warusfel03.pdf

So, can patents cover algorithms and their implementations? in Europe?
In France? I don't know. Is IPOL infringing the French law when
distributing the source code implementation od a US patent? I don't
know.

That's why my position would be to focus on our work (document good
algorithms and provide reliable implementations), and add a warning
when we know there may be a problem. If we try to do more, we can't do
it correctly and we will spend a lot of time and energy on something
we don't understand.

-- 
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/20110406/522565e8/attachment.pgp>