Fwd: Moving forward

[Jani Pirkola]

Hi realXtend,

this is my post to roboworld mailing list, FYI:

---------- Forwarded message ----------
From: Jani Pirkola <jani.p...@realxtend.org>
Date: 2008/9/5
Subject: Re: Moving forward
To: robo...@googlegroups.com

Dan, Ben,

I don't believe that GPL can catch if same developer looks at both GPL viewer code and BSD server code - they are using different languages, and copying code does not really even make sense from viewer to server - I wonder who would want to do that?

realXtend is still not mixing viewer and server developers because we would like to be able to contribute code back to OpenSim. And as we have enough developers, this does not hurt us too much although it is an annoyance.

This might change in the future, because DeepThink is doing the server side modularization work for realxtend, so that our features are in separate modules from OpenSim core.

The protocol and architecture discussions here, including also and even especially Dan's ideas, is seriously affecting realxtend future work. We are starting a task force to discuss that further, initially lead by MattiK - he will invite people to that (he should have done that already this week - Matti?) and the discussions and results will be made public.
We will also start to do a light weight client based on the work of that task force. The new client is going to be cross platform and extensible. Initially it is IM client to start with, but it will evolve over time to be more than that.

Best regards,
Jani

2008/8/29 dan miller <danb...@yahoo.com>

> What if, for initial prototyping purposes, one made an existent-enabled
> system that
>
> -- used OpenSim on the server
>
> -- used a very simple, newly-created client which **has no
> graphical interface**

I think this is similar to my suggestion of going with Opensim and Openviewer, both of which are BSD-licensed.  The trick would be to modularize the code so the Rex folks can take the existent stuff without getting involved in the graphics or the message layer.  Personally, I would find it hard to work on this without a 3D graphic representation.  My brain does not work on the purely abstract plane -- I need to visualize this stuff.

While Jani was supportive, he did say in one post that the existent idea was not going to make it into the next big bag of features.

I hope we can talk about this more next week.

-dan

[dan miller]

> this is my post to roboworld mailing list, FYI:

and here is my reply:

hi Jani --

All of this sounds reasonable. The fact remains that many in the Opensim group have an issue with even the appearance of GPL problems. The arguments for this extend to potential patent IP as well as copyrighted code. But further than that, there is a strong desire on the part of our community to have control of our own destiny, and that means weaning ourselves from the Linden teat, as it were.

For that and many other reasons, there seems to be momentum for the creation of a BSD-licensed client program, initially SL-compatible but eventually capable of supporting extensions to the SL protocol, including some of my concepts wrt distributed behavior/physics ("existants"). I am planning (hoping) to be a major contributor to this effort. We intend to implement it without having any direct access to Linden's viewer source, and we will probably impose similar restrictions on any potential contributors (as we do today with OpenSim).

So far, we have successfully wrapped the libomv libraries in Python (native CPython as well as IronPython), and I have written a (very!) rudimentary text-based client that can walk, run, fly, and chat on an Opensim/SL server. Our next step is to begin integration with Python-Ogre for the scene rendering.

The two big differentiators between this project and what I understand of Rex's plans at this time are:

1) BSD licensing -- both for Opensim's benefit, but also because we want to foster unlimited utilization of the codebase, including in mobile and embedded environments where GPL code is not practical

2) Cross-platform. From the get-go, this project will compile on all three major platforms -- Win32, OSX, and Linux.

There seems to be quite a bit of overlap between what we're thinking about and what Rex wants to achieve. Even if we don't share code, I believe we should start thinking about protocol compatibility. For instance, if we support mesh-based prims, it would seem sensible to understand how Rex has achieved this, and hopefully we can do it in such a way that our products are interoperable.

If it turns out that we are sufficiently aligned in our goals, it may make sense to combine forces -- but that will be a community decision when appropriate.

Daniel B. Miller
aka danx0r
life is a simulation

--- On Fri, 9/5/08, Jani Pirkola <jpir...@gmail.com> wrote: