[Opensim-dev] thinking about a viewer

[Mister Blue]

I'm interested in the 'renderer' as separate from the 'viewer' (made up of
renderer and components) and the architecture of the virtual world
infrastructure.

I like the move into the browser (all JavaScript and WebGL) to make a
viewer that is independent of any libraries or OS. But that said, there is
a lot to do with the infrastructure that is behind the virtual world.
 Ideally, any infrastructure should work with both virtual worlds and
augmented reality applications. So it should be about place and view and
interest.

Unity3D is a tech (and as a business they are interested in lockin) but, in
general, we should defined protocols between the viewer and the world
whether the world is SL or OS or whatever.

So, next steps, what forum should we have the architectural discussion?

-- mb


On Wed, Aug 13, 2014 at 9:37 PM, Diva Canto <diva at metaverseink.com> wrote:

> On 8/13/2014 7:33 PM, Mister Blue wrote:
>
>> Now, talking about Diva's challenge...
>>
>> The problem with creating such a solution is that the existing tools are
>> welded together -- 3d renderer is embedded into an app that presents a
>> pre-defined UI, chat system, camera system, ... It is all or nothing.
>>
>
> Assuming we'd develop a WebGL viewer more or less from scratch, this
> wouldn't be a problem [for the WebGL viewer]. We can separate protocol from
> rendering engine from [programmable] UI.
>
> Being able to use the Linden viewer as a collaborative authoring system
> for multi-user 3d content for the Web would be super cool. [Note to patent
> trolls lurking on this list: I said it here first]
>
>
>  Systems like RealXtend try to solve the changeability problem by making
>> everything programmable. So there, an author has to program and deal with
>> the interaction of all the parts.
>>
>> What is missing is an 'authoring language'. What is the 'language' (and I
>> use that term for a solution loosely) that a scene author can use to create
>> the challenge scene. Technically, the parts should be separate and
>> composable. From the author's point of view, the world should only consist
>> of the actors and the interactions desired.
>>
>> There are scenario editing and interaction 'languages' available. The
>> Unity graphical state system is cool. I'd like to see others that were more
>> scene or world oriented. The underlying tech would be of composible pieces
>> that supported the construction of the desired world.
>>
>
> Yes, Unity3D is definitely a good reference point. I find it lacking on
> only two things that I think are superior in the Linden system: (1) ease of
> creation: prims are easier for simple things; (2) collaborative authoring:
> creating a complex scene/game with multiple developers in Unity3d is a pain
> in the neck. So, we could borrow a few ideas from Unity3D and add them to
> the Linden viewer -- things like specifying camera movements, controller
> objects, 2D UI, etc. Even if these things are not runnable in the Linden
> viewer itself, they would run in the WebGL viewer. So while developing, we
> could run both the Linden viewer and a web browser pointing at the sim.
>
> I want this to happen. I just don't know anything about rendering, so I'm
> useless wrt developing a WebGL renderer, and I'm waiting for the hero who
> does that in open source. But you can count on me (and I'm sure others
> here) to jump in on everything else -- protocol, authoring language, etc.
>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at opensimulator.org
> http://opensimulator.org/cgi-bin/mailman/listinfo/opensim-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20140815/7a87d407/attachment.html>