Linden Lab on CatB (Re: [sldev] Voice=Proprietary)

[Rob Lanphier]

On 3/13/07 8:15 AM, Boroondas Gupte wrote:
> That would be the old question of the cathedral and the bazaar
> <http://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar>. Linden
> Lab will have to make clear where they see their open source efforts
> on the scale between those extreme paradigms. They are of course free
> to redefine that for every single part of the project. I has not to be
> the same for the entire thing and also might change over time. Anyhow
> I think it would help community developers a lot if the status quo
> concerning this was known as well as LL's plans about where to go from
> there.
>
> I don't know if LL has worked this out for themselves, yet. If not
> they should now ;-)
If anyone here ever needed proof that Linden Lab is not a monolith, look
at Dave's response.  It's a very frank and accurate assessment of where
we are today, and I'm really glad he sent it.  However, I'm going to
give a different answer, because I "gotta be me", and while I agree with
most of it, there's one or two points we differ on.  I also love the
fact that he volunteered to be "bad cop" possibly without knowing it.  :)

What I believe is unprecedented about our source code release is that we
did it at a time when we didn't need the press (lord knows, we didn't
need the press), and while our popularity was (and is) still very much
on the rise.  We didn't set out as an underdog disrupting an existing
market or as a has-been trying to recapture the magic.  The result has
been that outside developers have been instantly attracted to it...lots
of patch submissions, lots of bug reports, tons of developer list email
all within the first couple of months.  This is not the typical response
to corporate open source initiatives, but we're not complaining.

Let's play through a more typical script first.  A company loses their
market leading position -- and they find themselves scrambling to
recapture the magic.  As a tactic for reclaiming their market
leadership, they open up the source code.  Then they discover that
instead of having a proprietary product that few people are interested
in, they have an open source product that few people are interested in. 
Worse, they have a lot of entrenched, crusty development processes that
weren't really working for them before that they try to export to the
world.  If they are lucky, the free software community will tell them
exactly what they are doing wrong, and what they need to change ASAP to
fix what is clearly broken.  If they are smart, they listen.

Our situation is different.  I know our development process isn't
perfect, but it's working for us now.  Even if I thought our development
process was broken beyond repair, I'd have a tough time making the case
to everyone else that that's true.  As a relative newcomer to a
successful company, it's not as though I should barge in and say
"everything you guys have been doing is wrong...we're changing
everything from top to bottom".  I believe that free software
development processes represent a big improvement to how we build
software.  I believe that we should be more and more stringent about the
licensing of the technology we incorporate in our product.  However, I
don't think it's smart for us to drop everything we're doing and
instantly transform the way that we work on a day-to-day basis, and I
suspect that if I did run around like a bull in a china shop telling
everyone here how totally screwed up their way of developing software
was, I'd be roundly ignored and highly ineffective (and thus, soon out
of job).

So, we're taking it easy, trying to be careful not to kill the goose
that laid the golden egg.  I'm patiently explaining why some of the
things we've traditionally done probably aren't going to work for us
forever.  I'm explaining how things can be better, and doing my best to
ensure that for the pain we endure, there's actually corresponding
gain.  The incremental improvement model may not be fast enough for
everyone's taste here, but it's how things have always been done at
Linden Lab, and I see no good reason to break with tradition in that regard.

You all can help with the education process.  I read "The Cathedral and
the Bazaar" when ESR first posted his presentation, but I can't say for
sure if everyone at Linden Lab has (Lindens in the crowd, take
note...).  We appreciate these type of pointers at relevant times.  As
Dave pointed out, you're not going to get a "policy statement" on behalf
of "Linden Research, Inc." regarding where we fall on the CatB
continuum, because even if we published it, there's no guarantee the
developers here would read it anyway, so it wouldn't mean much. 
Patient, reasoned explanations of why a more open approach is more
appropriate for our situation will be far more effective than some of
the ultimatums and demands that others have sent to this list.

With regards to a Subversion repository (not brought up specifically by
Boroondas, but elsewhere on the thread), it's something we're working
on...as in, you may notice that "svn.secondlife.com" actually resolves
to something.  There's nothing there yet...we're still working out how
we're going to use the thing, and the more time I spend writing emails
like this, the less time I spend working out that problem.  However,
we're gradually getting all of the trappings of a "real" open source
project.  Please be patient.

Thanks
Rob


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 249 bytes
Desc: OpenPGP digital signature
Url : http://lists.secondlife.com/pipermail/sldev/attachments/20070313/eb76ab91/signature.pgp