Re: [Bino-list] Bino licencing problems...

[Martin Lambers]

Hi Vittorio!

On Mon, 20 Feb 2012 19:59:42 +0100, Vittorio Giovara wrote:
> I wanted to start a difficult discussion as it's a particularly "hot"
> thread, especially for an opensource project.

Well, at some point such a discussion is inevitable, so why not discuss
it now.

> I feel that using GPLv3 is not right for Bino and I'm going to
> explain why.
> 
> GPLv3 is a major step forward from GPLv2 and it takes care of a lot
> intellectual property loopholes that sneaked in previous versions.

> However one of the "new entries" is a nasty requirement over patent
> licenses: in layman's term, if you own a patent and implement the
> feature under a gplv3 license, you automatically give a *free*,
> *perpetual* license to *everyone* and every derivative work.

This is not at all nasty: it makes perfect sense! How else would free
software be usable in jurisdictions where software patents are
enforcable?

> Let me make the devil's attorney and state that this does not bode
> well for a "new" (or rather "reborn") research field such as
> stereoscopy. In general it is very likely that newer technologies
> like depthmasks, 2D+delta, MVC and so on are being covered by one or
> more patents.

Maybe. But I don't think the patent problem in the field of stereoscopic
video is worse than in any other field of software.

> What does that mean for Bino while using a GPLv3 license?

> Well, the owners of such patents will never agree to submit any of
> their code under such licensing terms;

Probably true. But that is also true for holders of patents related to
video codecs, see for example the "Patent Mini-FAQ" at
<http://ffmpeg.org/legal.html>. So video players in general are
affected by this, whether they support stereoscopic videos or not.

FFmpeg, VLC, mplayer and others are successful in providing first-class
video codec technology without code submissions from patent holders, and
regardless of patent problems that may or may not affect some
jurisdictions.

I see no reason why Bino would depend on code submissions from patent
owners.

> moreover if any contributor
> implements anything that might be in a patent, he or she
> automatically grant a license for something that they doesn't posses.
> In other words,  this makes them and Bino infringe that patent.

This is not correct. Sec. 11 of the GPL is explicitly limited to
"patent claims owned or controlled by the contributor". (Anything
else would very likely not be a valid license term anyway.)

> However there are more liberal licenses, such as the BSD or MIT
> license, that basically state "let us code together and not deal with
> the rest": if Bino were to switch licenses, I believe that we could
> continue working and using Bino like now, but with the possibility of
> having more and more features added without risking of violating
> anything. Besides I think it could attract new contributions from the
> industry which is usually scared away by such strict licenses.

The BSD or MIT license might encourage people to modify and repackage
the player, but I doubt that many modifications would be contributed
back to under a free software license. Why should they?

I prefer a license that requires modifications to be contributed back
once the contributor decides to publish its modified version.

This requirement may not always suitable for software libraries, but
it is perfectly reasonable for applications.

> Is there a strong reason for licensing Bino as GPLv3?
> Would you consider changing licenses for Bino? Martin, I would
> really like to know what do you think of this possibility and if the
> group shares any position on this idea.

A license change requires convincing arguments.

In your mail, you state that a license change would
1. make life easier for companies that threaten users and developers
with software patent law suits, and
2. allow companies to keep modifications to themselves instead of
sharing them.

Neither argument seems convincing to me.

Regards,
Martin