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

[Rob Lanphier]

Hi Andrew,

On Thu, 2003-12-11 at 15:55, Andrew Sobala wrote:
> Hi Rob,
> 
> On Thu, 2003-12-11 at 23:26, Rob Lanphier wrote:
> > On Thu, 2003-12-11 at 02:59, Thomas Vander Stichele wrote:
> > > I want the users of my framework to have access to them, not myself.
> > 
> > I guess if you want users of GStreamer to have access to RealAudio and
> > RealVideo, you'll need to go through the same expensive combination of
> > licensing and R&D that we did.  
> > 
> > You might be able to convince us to come up with a license for use with
> > GStreamer, but I'm not optimistic.  The challenge is that RealNetworks
> > wants the industry to standardize on Helix, and uses RealAudio and
> > RealVideo as incentives to move to Helix.
> 
> I'm seeing a catch-22. Can you explain why this is wrong?
> 
> 1) RealAudio/Video is not licensed such that it can be used in other
> open source projects
> 
> therefore 2) The RealAudio/Video codecs are in a non-Free license
> 
> so 3) why should the community help if they cannot use these codecs in a
> completely Free library stack?
> 
> The thing is, if the codecs aren't Free to modify, then Helix when using
> these codecs is not fully Free either. So basing anything on Helix means
> we end up using a non-Free media platform. So people won't/can't
> contribute to it.

Let's dissect that last paragraph very carefully:
*  The thing is, if the codecs aren't Free to modify, then Helix when
using these codecs is not fully Free either.

True statement.

* So basing anything on Helix means we end up using a non-Free media
platform.

False.  Helix is codec-agnostic.  There are already fully free decoder
implementations of Ogg Vorbis, SMIL 2.0, H.263 video, and partial AVI
support.  You have a protocol stack which is a very robust RTSP/RTP
implementation, which has been through several rounds of IMTC
compatibility testing for use with 3GPP systems.  We're adding new
codecs all of the time, and would love to see more open source codecs
added to the system.

As an example of a presentation that isn't possible in another system
(that I'm aware of), but is fully open source, check out this demo
(played in the Helix Player):
http://rtsp.org/2003/demos/oggsmil/oggdemo.smil
Tarball:
http://rtsp.org/2003/demos/oggsmil.tar.gz

The presentation is Ogg Vorbis audio, and JPEG slides exported from
OpenOffice.  The timing is done using SMIL 2.0, with great looking
transition effects between the slides.  *Everything in this presentation
is open source*.  Even if you don't bother downloading and installing
the Helix Player, at least look at the text of the SMIL file above. 
It's pretty easy to read the XML and figure out what's going on.

So, I understand your interest in a fully-free platform.  What I'm
saying is "great...focus the efforts on making Helix into a fully free
platform".

> > That's the bad news.  The good news is that we're willing to invest
> > heavily in Helix, and plan to build up a large industry-wide ecosystem
> > on top of Helix.  It already is a great hybrid of open source and
> > commercial licensing, and the more successful we are with it, the more
> > we'll invest.
> 
> You're using the word "commercial" where you should be using
> "proprietary", it's quite possible to build a business model
> ("commercial") around open source licenses (see Red Hat, Sun). Just a
> linguistic point :)

That's a fair point, but I should point something out.

I believe the bulk of Sun's revenues are from their proprietary software
businesses.  Red Hat has a smaller gross revenue than RealNetworks, and
Red Hat is the 800-pound gorilla of pure-play open source businesses. 
All of the other big open source businesses that I'm aware of are
hybrids, and the proprietary bits are where the margins are.

That said, I agree that open source business models are the future (my
personal opinion).  It's just that they aren't the present.  We're
putting in a good faith effort to make a solid open source business, but
we have to deal with the current realities.

> > > How hard can it be to
> > > a) make binary-only codecs freely available without clickthrough license
> > 
> > Very.  We have very specific licenses to our codecs, and we're
> > contractually bound to specific terms.  Moreover, as stated above, we
> > don't make money on codecs.  We *pay* money for codecs.  We make money
> > off of Helix, so we want people to use Helix.
> > 
> > > b) write a GPL/BSD/Real License/... API and library for using them
> > 
> > So that you don't have to use Helix?  And why would we do that if our
> > goal is for you to use Helix?
> 
> I think, if Helix is a super-master-class media library, we'll end up
> using it. If the whole thing is Free. (We're talking deep-future here,
> of course.)
> 
> If gstreamer trumps helix, we'll end up using gstreamer. In Open Source
> software, the way to get people to buy-in to the technology is by being
> the best. Holding the codecs hostage to a proprietary license where they
> can only be used by helix won't get people to develop helix because
> writing software that _avoids_ the use of proprietary licenses is what
> keeps them ticking.

Fair enough, but I'm not sure how GStreamer really avoids proprietary
codecs.  Can you explain?

Rob
-- 
Rob Lanphier, Helix Community Coordinator - RealNetworks
http://helixcommunity.org http://rtsp.org http://realnetworks.com