[sldev] Voice=Proprietary

[David Fries]

On Tue, Mar 13, 2007 at 09:26:35AM -0700, Dave Parks wrote:
> We're very serious about keeping Second Life open, not just in source 
> form, but as a platform.  We have not to this date and never will do 
> anything that prevents a class of people from accessing Second Life.  
> Any outside service we rely on, any proprietary library that we license, 
> any hardware that we interface with will all adhere to an open 
> standard.  No exclusive deals, no monopolistic endeavors, no artificial 
> limitations.

What open standard would that be?  That and when proprietary and open
are in the same sentence?  Hardware?  Are there some hardware products
being designed to work with Second Life?

While I'll have to agree that being without voice does not prevent a
class of people from connecting to Second Life, I do wonder what kind
of an experience one without the voice chat will have when the greater
Second Life is voice enabled.  Will people be barred from education,
or social events because they can't hear what is going on?  That's not
under Linden Lab's control, but those who have voice chat could
effectively make it that way.

I would say that voice is the coming third proprietary challenge to
the Second Life viewer, and this is what I've seen of the first two.

In the case of jpeg2000 decoding there were two libraries available
when the source was released, one proprietary, the other open.  The
proprietary one works (but no source, not re-distributable, limited
platforms), OpenJPEG is slow, crashes, and looks bad.  That's the
current state of the code as released today.  The community (I happened
to be the one to put things together, but others provided some
pointers and code to do it), fixed OpenJPEG and it should be included
in the source release any time now.  But Linden Labs put the OpenJPEG
support in there in the first part, there were just some problems
left.

You could connect to Second Life without either library, and indeed
at least one has to keep it from crashing when they didn't know about
the fixes.  I've seen Second Life without textures (due to another bug
to track down), it's not completely gray some things to have color,
but it isn't pretty.

There is the current environmental sounds fmod.  You can't get source
code to that.  But the codec is Ogg Vorbis, I don't see any reason
another sound system couldn't be built to replace it.  But this isn't
the same as voice chat.  Sure you loose something of the experience,
but as I've been connecting without sound since before I had gotten
OpenJPEG up and running I'm not sure what I'm missing anymore.

That brings us back to voice chat.  Will it go the way of jpeg2000,
pretty much required and eventually supported in open source, or
environmental sound, it would be nice to have but if you don't you
aren't missing anything important?

> B) Decide that the technology you've been given is not good, and try
> to compete with it.

Could those voice servers support more than one audio codec?  They are
encoding and decoding a unique stream per user, so I don't see why
they couldn't exchange Siren14 with one user and Speex with another.
That would let the open source community write the client side with an
open codec.

With a closed codec the only legal solution I could think of is to
create another separate voice system.  One that people couldn't talk
to people using the proprietary system.  Would Linden Labs like a
solution B where people are split up into groups that could only talk
with those who have the same solution?

Just think of how far the cell phone network would have gotten if you
could only talk to people who also had cell phones.

-- 
David Fries <david at fries.net>
http://fries.net/~david/ (PGP encryption key available)