[vwrap] Browser Based Virtual World Viewers [ Was: Re: Comments on http://tools.ietf.org/html/draft-ietf-vwrap-intro-00]

[Meadhbh Hamrick <[e-mail]>]

oh and heck. i forgot i was going to comment about your comment about
browser based VW viewers. moving it to a new thread...

 Alzo Spracht Cristina Videira Lopes:

> Personally, I'm becoming more and more interested in web-browser-based
> viewers, or something like them. I.e. viewers that are simply shells that
> receive, essentially, a lambda implementing the viewer itself. That's the
> way to go. But that doesn't exist yet, not in production at least.

yeah. while i'm not an unqualified fan of jibe/unity-3d, i think it's
come a long way. i understand why reactiongrid built jibe on top of
unity and i understand why unity really likes flash.

however...

HTML5 looks like it's going to be teh ossm. i'm a little sore that
WebGL emerged without a retained mode, and mildly sore that peeps seem
to be more interested in implementing websockets than XMPP in modern
browsers. but oh well. tempora mutantur, nos et mutamur in illis.
there are still some very cool things coming down the pike.

a. better / more ubiquitous support of WebGL.- for rendering 3d scenes
in your browser.
b. HTML5 AUDIO support for streaming stereo audio - so you can have
something like 3d audio (right in your browser! w00t!)
c. HTML5 DEVICE support for microphones and joysticks - so you can use
your headset to talk into the virtual world as your favorite game
controller will work via your browser.
d. ubiquitous support for websocket - so we won't have to rely on long
poll for the VWRAP event queue.

there are still some challenges:

1. it's not clear that web browsers will be able to process event
updates as fast as a dedicated viewer.
2. there's still a little bit of uncertainty re: accessibility for
WebGL experiences in browsers (though it looks like peeps are working
on this..)
3. it introduces new security issues. using the same browser to access
your bank website as an online 3d world. we have to make sure any vw
software deployed through the browser does not introduce the ability
for bad guys in the virtual world to access your bank account.

and an observation or two...

* i think by the time the VWRAP suite is fully specified, web browsers
will be sufficiently capable of rendering at least low quality virtual
worlds.
* i think the challenges listed above are not unsurmountable.

and most importantly...

* it's entirely possible to build a separate viewer application that
uses the mozilla or webkit rendering libraries, but is NOT a web
browser.
* an application like this, that can control the cache behavior,
possibly include a plugin for faster handling of object updates, etc.
could be a great way to begin working with web based world clients,
but without having to worry about differences between web browsers.

and i think it's important for this list, because...

* unless anyone objects strenuously, i'm going to add a paragraph or
two in the intro doc explicitly calling out web based viewers as a use
case (but not the only use case) for VWRAP virtual world clients.

-cheers
-meadhbh