Re: [gst-devel] Re: Helix Player virtual team meeting



Hi Rob,

I'm not really an expert in open source licensing, but I thought I'd
ramble into this thread. I'm hoping to represent some of the views of
the FOSS community to you where I think you may misunderstand them; of
course, I may not be accurately representing the views at all! But I'm
going to give it a go.

On Wed, 2003-12-10 at 23:57, Rob Lanphier wrote:
> On Wed, 2003-12-10 at 13:17, Thomas Vander Stichele wrote:
> > Reading that part again and again, it still seems to say that all
> > copyright of anything contributed by people is assigned to real.  If
> > this is not the case then please tell us what it is I am misreading, but
> > quoting other parts of the license doesn't really change the fact that
> > this part of the license seems to say so.  That is also what other
> > developers I talked with seem to get out of it.  
> 
> We didn't want to be in a situation where someone could fork the code
> without us having some recourse.  We're not worried so much about
> small-time contributor in their basement, but it would really be bad for
> some large, well-funded contributor to strip off our commercial terms
> (the RCSL), and use our codebase to compete against us.
> 
> I'm not sure I see how this hurts a potential contributor.  Yes, we
> could use their code in a proprietary product, but it's not like the BSD
> situation where you can have total parasites who take from the
> community, and never give back.  This is a special right granted to us
> as contributors of the code....it seems like reasonable quid pro quo to
> me.

The quid pro pro is this:

1) RN makes code open source, RN gets improvements back on the code.
2) FOSS people get FOSS codecs, FOSS people improve RN's code. See 1.

The thing is, what makes people tick in open source projects is creating
a good Free [1] project where people can use it, change it, and
redistribute it. If their code can be reintegrated into a proprietary
product, then they run the risk of all the company's development going
into the proprietary product, the company also integrating the
community's work into the proprietary product, and the open source
product just not getting critical mass. This is only a problem where a
single company is a large contributor to a product, but that could well
be the case with realplayer/realone/helixcode/helix community/real
codecs.

> > I will mail the gnome board as well asking them if they can get someone
> > more qualified to translate the legalese for us.  But if you want to win
> > us over I guess you should start by explaining why this bit of the
> > license should be intrepreted differently.
> 
> I'd be disappointed if a potential contributor turned away because they
> dwell too hard on the rights RealNetworks gets rather than focusing on
> the rights the contributor gets.  We've invested millions of dollars in
> this codebase, and plan on investing millions more, so it seems rather
> stingy to begrudge us a few extra rights.

Don't think of the whole codebase. Think of a single contribution.

The copyright on that contribution is originally fully owned by the
contributor: the decision they have to make is how to license it. Given
that many contributors will be doing this in their spare time in order
to advance Free Software, they may find it quite reasonable not to want
to license it in a way that allows a company to sell it in a proprietary
product (as opposed to selling the Free/Open-Source product).

No-one minds Real keeping the copyright on their contributions. Its
demanding that other people hack for free for what they can turn into a
proprietary product that people may object to.

> 
> > > So, regardless of the fact that RealNetworks, as the Licensor, can 
> > > create proprietary applications, you have the right to continue to use 
> > > the open source version.
> > > 
> > > There are plenty of examples of where this is happening with GPL 
> > > software today.  For example, much of the code that Raph Levien wrote 
> > > was facilitated by this funding model (http://www.artofcode.com/)
> > 
> > Right, so what you're saying is, every version that is put out under
> > this license will always be usable as opensource and available.  That is
> > in fact reasonable.
> > 
> > It also somehow implies that there might be a time when newer versions
> > will not be opensource or easily available free.  Which might be where
> > some people have problems.
> 
> This can happen even with GPL.  Example: SourceForge.  VA Software took
> a package that they licensed as open source, and made a proprietary
> product out of it.
> 
> We're just being explicit about our intentions.  We hope that the
> contribution from the community will be significant enough to never make
> us want to do that, but we owe it to our shareholders to keep our
> options open.

When Sourceforge made a proprietary product out of their software, it
would be a breach of their licensing terms if they included third-party
improvements to which they did not own the copyright. I'm guessing that
there was very little community around the software (as opposed the
site) making VA Software effectively the only contributor so they could
easily proprietarise their code.

However, if you want to build up a community, then you do have to take
into account what the people who write the code want to happen to it. A
lot of them will not want to donate it to a proprietary product.

Remember, companies essentially get open source work done by the
community "for free." People spend time hacking on things because they
enjoy the creation of a Free product. They may want to protect their
code from uses in non-Free software.

> > > > The GStreamer team is pushing hard to get an integrated and
> > > > distributable media framework and we are currently trying to fund the
> > > > development of Theora codec so that we can get a completely free and
> > > > open solution for multimedia streaming/playback.
> > > 
> > > Actually, RealNetworks has funded the Xiph Foundation in the past:
> > > http://www.realnetworks.com/company/press/releases/2002/xiph.html
> > 
> > I read the press release back then, I reread it right now.  It never
> > offered specifics, but as far as I can make out it just entailed getting
> > a Vorbis plugin written for Helix, right ? Which is all fine, but since
> > the code for vorbis was out there under a well accepted open source
> > license, it could have been done by Real itself.  What I'm saying is, it
> > didn't really help the development of Vorbis, though I can see why it
> > makes sense for Real to pay codec developers to develop plugins for
> > them.
> 
> Look for more announcements on this front.
> 
> > > RealNetworks has a long history of being as open as the business climate 
> > > will prudently allow, and I don't see any reason why this won't continue.
> > 
> > For me, and for a lot of other developers, the success of Helix, and our
> > possible involvement in it, stands or falls with your license, what it
> > intends to say, and what it legally says.  Most developers seem to get
> > from the license that all copyright is completely transfered to real.  I
> > think you know why this is not acceptable.
> > 
> > So possible routes, if this matters to Real, are:
> > a) change the license
> > b) explain to us what it is that we understand in the wrong way
> > c) go right ahead without us
> 
> It very much matters to us.  We'll consider (a), but a very strong case
> needs to be made for that.  It helps to finally be having these
> conversations on a helixcommunity.org list.
> 
> However, it's important that we explore (b) (rephrased: come to a
> negotiated set of common goals).  Competition is always good, but I
> think we're getting ample competition right now from other completely
> proprietary frameworks.  
> 
> > Also, from an outside view, it is not very clear how open Real has been
> > in the past.  Looking at the actual situation, the real codecs are one
> > of the last to not have been reverse-engineered in some project.  The
> > least that could have happened is some opening of codecs well beyond
> > their expiration date.  It seems to me that your sentence translates to
> > "the business climate doesn't allow Real to be open in a way that is
> > useful to us".  While you may disagree with that interpretation, you
> > should realize it is one that many share and that you will have to
> > dispell :)
> 
> The fact that you've reverse-engineered some codecs doesn't mean you've
> reverse-engineered the patents out of the U.S. legal system.  Do all of
> these reverse-engineered codecs come with a patent license?
> 
> We've actually got patent licenses for the codecs we distribute.  While
> that may not be useful to open source software projects that willfully
> ignore US patent law, it makes a big difference in the aforementioned
> business climate that RealNetworks has to operate in.

Open source projects care a lot about patent law. Most members of them
don't like it :-). This is because they prevent open source projects
interoperating with their proprietary cousins.

What is the Real policy on patents? The "ideal" answer, in terms of
being friendly to the community, is to grant users and developers of
FOSS a royalty-free license to use certain Real patents. This is what
Red Hat has done [2]. I'm curious to know what the Grand Plan is as far
as Real is concerned.

> > That is not to say Real is completely closed; but you must understand
> > that as free software developers we are wary of being used as cheap
> > labour on a project that is less free than another with the same scope.
> 
> I understand your concern.  We'll 'fess up to being a big stupid company
> who doesn't fully understand why people contribute to open source.  In
> aggregate, we're guilty as charged.
> 
> However, we're not asking you to contribute to RealAudio and RealVideo. 
> We're asking to collaborate on a media framework.  It so happens that as
> a bonus for using our media framework, you get *legal* access to
> commercial-grade codecs that are available today, as well as access to a
> number of open source codecs and datatypes (including Ogg Vorbis and
> SMIL 2.0).  We plan to add more, and we welcome the contribution of
> others.

The thing is, the projects that would use a Real codec will very
probably be LGPL or GPLd. These are clearly Free licenses.

But as hackers, a lot of us want our program *and all its dependencies*
to be Free so the whole thing can be modified, and given to our friend
across the road without breaking any laws. The question is whether the
RPSL is sufficiently Free or not. And whether the license is Free is the
showstopper on whether people will contribute back to the codecs or the
media framework.

-- 
Andrew Sobala <aes gnome org>

[1] http://www.fsf.org/philosophy/free-sw.html
[2] http://www.redhat.com/legal/patent_policy.html

All opinions in this e-mail are hopefully my own and don't necessarily
represent the opinions of the GNOME foundation, etc, etc.

Attachment: signature.asc
Description: This is a digitally signed message part



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