[whatwg] Re: XHTML + X3D (was Re: Web Applications and 3D)

[Ian Hickson]

On Fri, 29 Apr 2005, Matthew Raymond wrote:
> 
> I'm not suggesting integration with HTML5. I'm perfectly happy with 
> using X3D (or some other separate language, like XGL). My concern is the 
> lack of analysis of how these technologies fit together. It's like 
> fitting an engine to an airframe. I don't want the engine to necessarily 
> be part of the airframe design, but they still have to work together, 
> and you won't know where the problems are until you try fitting them 
> together.

I guess I don't understand what you mean by "fit together" if you don't 
mean "integrate". Any format can "fit" with XHTML, including PDF, Flash, 
PNG, SVG, MP3, AVI, MPEG, etc, and including X3D and its predecessor VRML, 
if integration isn't the point.


> > > There's also the potential to use XBL to apply special behavior to 
> > > specific tags. For instance, instead of having a static Sun in the 
> > > sky, you could have it rise and set by using XBL to pull in code 
> > > that moves the model and lighting around.
> > 
> > What sky? :-) I haven't seen many applications with skies...
> 
> I'll try to come up with a less game-like example next time...

Would you really expect a game to be written with declarative 3D? I guess 
it's possible, but my impression based on talking with 3D people and game 
designers asking for Web-deployed 3D games that re-use the existing 
framework (HTML) is that they really just want access to the underlying 
APIs so that they can make their own, optimised 3D engines that do what 
_they_ want, as opposed to having a particular framework.

(Note that DOM manipulation is relatively expensive, by the way, in terms 
of runtime performance.)


> Question: Game != Application?

No, not at all -- games are in fact often the leading edge of 
applications, and often have even more extreme needs. Games are 
definitiely part of the WHATWG target application segment.



> > The biggest problem with integrating X3D and XHTML is, as far as I can 
> > tell, the lack of a namespace to identify X3D content. (This is 
> > similar to the problem preventing Docbook integration.)
> 
> Good, now we're getting somewhere. Let's say X3D did have a namespace, 
> like "x3d". Could we use <x3d:x3d> or <x3d:scene> as the 
> container/viewport?

Assuming X3D defined how that worked, yes. (It would basically need to 
define a the contents of, and dimensions of, a 2D rectangular area that 
represented the X3D content. That would then be a replaced element and 
would fit into the CSS model directly.)


> > If you want declarative 3D integrated into the Web browser pipeline, 
> > alongside the DOM, JS, XHTML, SVG, CSS, XBL, etc, I encourage you to 
> > bring this up with the Web3D consortium.
>
> Very well. I'll also look into competing languages being developed.

Cool. (Note that X3D is on the ISO standards track, which gives it a lot 
of weight.)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'