[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Royalty free gstreamer plug-in

Thomas Vander Stichele wrote:
> > That is not accurate.  Use of the MIT license doesn't change the patent
> > situation.
> Could you expand on that point a bit ? Calling something "the patent
> situation" doesn't tell me exactly what you mean.
By "the patent situation", I mean that MP3 is covered by patents, and
that to use or redistribute any technology that takes advantage of a
patented technique, you must obtain a license from the patent holder.
> FWIW, for most distributions the problem with mp3 code is precisely that
> all these decoders were GPL-licensed, and the GPL has a clause that
> terminates your right to distribute if you cannot distribute freely to
> all (remember, IANAL).  So the problem was in the GPL, not in the code
> or in the fact that patents apply.
> Feel free to correct me if I'm missing something.
That is only part of it.  The other part is that we are in the United
States, where use of patented techniques without a license is illegal. 
Your licensing means we can redistribute from a copyright perspective,
but we are still constrained by the patent.  Under the GPL, we lose the
copyright if the codebase is impacted by a patent.  With the MIT
license, we still have a copyright license even though the technology is
covered by a patent, but it is still illegal for us to use and
redistribute the technology without a patent license.
> >   If patents were so easily defeated, it would have happened
> > long ago.
> This does not defeat patents.  This only aims at achieving a very
> specific thing: allowing distros to ship both source code for a decoder,
> without violating the source code license, and binaries built from it,
> with the necessary patent license attached.
The point is that a contract is still needed for the binaries.  We could
have gotten a patent license before, but we will not be given a patent
license that automatically covers us and all downstream users and
distributors.  Thanks to you, we have something we can distribute from a
copyright perspective, but not a patent perspective.
> One of the reasons this never happened before is because nobody wanted
> to invest in creating a MIT codebase to do so (remember - this means any
> other company can "rip Fluendo off" and repackage their code) - all free
> software decoders that I know of are under the GPL.
This is something unique, and parts of the community can thank you for
it.  Unfortunately, it doesn't help Fedora or others who still have to
worry about the patent constraints.
> >   None of the licenses that Fluendo has will cover
> > redistribution by others.
> Are you sure ? Fluendo offers any distributor the possibility of signing
> a contract that allows distributors to rebuild binaries for which the
> patent license Fluendo negotiated is transferable.
That contract is a killer.  Also, it wouldn't make much sense for us to
include a plugin that can only be used with other non-GPL software. 
Most of the software in Fedora that uses the gstreamer backend is GPL
software.  This solution is GPL-incompatible, and that causes a lot of
problems that are hard to overlook.
> > That might work for RHEL, but it is meaningless for Fedora, and
> > basically a non-starter for any non-commercial redistribution who's
> > looking for the legal way out.  Even if it were included in RHEL, it
> > would be much like the Java components that cannot be redistributed with
> > the rest of the open source components.
> > Lacking language about patents doesn't take away the liability.  Quite
> > the opposite, in fact.  Using the MIT license is telling everyone else
> > "Hey, if you get sued for using this, remember - our license didn't say
> > anything about patents, and it didn't offer you any protections."
> The point is that "use" of a piece of software applies to the binary,
> not the source code.  You cannot get sued for having a piece of software
> code.  You can get sued for distributing a working (in this case,
> binary) version of the code without the permission to do so, or for
> using a binary version for which you do not have a permission.  Of
> course, IANAL, but if I'm wrong about this, please argue the specific
> points I'm wrong about.
> In this respect, any piece of code currently shipped by Fedora could be
> the subject of a lawsuit for any of the customers.  Fedora is not
> guaranteeing their users they won't get sued for anything in the distro,
> so I don't see why it would be different for this particular package.
Fedora isn't making any guarantees, but we are trying hard to protect
our users and downstream distributors.  NTFS is a similar situation. 
http://fedoraproject.org/wiki/ForbiddenItems isn't just for us, it is
for us, our users, and downstream distributors and builders.
> Thanks for your insightful comments, I hope to form a clearer picture
> for myself on the situation.
> Thomas
> Dave/Dina : future TV today ! - http://www.davedina.org/
> <-*- thomas (dot) apestaart (dot) org -*->
> Is there a voice unkind
> in the back of your mind
> saying "maybe you didn't know him at all"
> <-*- thomas (at) apestaart (dot) org -*->
> URGent, best radio on the net - 24/7 ! - http://urgent.fm/
What you're doing and what you're trying to do is great, but there is a
distinct reason that MP3 has been such a problem in the past.  You've
made some progress, but we are still a long way away from being where we
need to be in order to have free MP3.  The patent holder doesn't want
their technology in the wild, and until they are ready to release their
hold, that situation won't change.  Many thanks, congratulations, and
good work, but there's still a lot that has to happen before Fedora can
have MP3 support.

Patrick "The N-Man" Barnes
nman64 n-man com

Have I been helpful?  Rate my assistance!  http://rate.affero.net/nman64/

Attachment: signature.asc
Description: OpenPGP digital signature

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]