[opensource-dev] oh give me a break

Gareth Nelson gareth at garethnelson.com
Tue Mar 16 02:49:09 PDT 2010


On Sun, Mar 14, 2010 at 9:28 PM, Soft Linden <soft at lindenlab.com> wrote:

> We'd need to provide a way to move off of Havok while still remaining stable with
> insane physics content,
No you wouldn't, if you wanted to release the server code with the aim
of increasing compatibility at the protocol level, the simplest
solution is to just dump all the code you can without including havok.
Then let the community do the work of porting to ODE or some other
engine. On SL itself, ODE may not be 100% compatible with existing
content, but on other grids it should be fine.

> deal with a lot of licensing issues
I have a simple solution to this one: if you don't 100% own the
copyright in (or have a suitable FLOSS license for) any one piece of
code, don't release it - just leave a big hole marked "code
replacement here"

> rework the server protocol to deal with untrusted peers and survive wider
> version differences, find a way to preserve the economy and creator
> rights, on and on.
releasing the code and opening the grid for external connections are 2
different things, though I suspect that for connecting untrusted
servers there are solutions such as OGP or opensim's hypergrid - i'd
even plug the litesim supergrid here if it still existed (one thing to
note about VW hosting startups: don't underestimate the resources
needed on your backend - got plenty of profit margin on individual
regions, but was slaughtered by backend resource usage)

> Even open sourcing the viewer was a huge time and
> resource investment, done with the calculation that the time invested
> would eventually pay off.
and it did - I suspect that the real reason you don't want to release
the server code is to maintain competitive advantage. That's fine, but
I just wish LL would be more honest and not claim it's about licensing
or other irrelevant matters.

One thing to note is that opensim is rapidly becoming a real
alternative to SL at least in terms of features and so the competitive
advantage of having the server code secret may not last long. The
danger to LL is that opensim et al will move more and more away from
SL compatibility while offering more and more seductive features that
draw away end users.

I don't know if anyone has done it yet, but a P2P opensim (real P2P -
where every client is also a server and they all sync the physics with
each other) will absolutely crush LL for a lot of "social" uses.

-- 
“Lanie, I’m going to print more printers. Lots more printers. One for
everyone. That’s worth going to jail for. That’s worth anything.” -
Printcrime by Cory Doctrow

Please avoid sending me Word or PowerPoint attachments.
See http://www.gnu.org/philosophy/no-word-attachments.html


More information about the opensource-dev mailing list