OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: XML multimedia specs -- help for the bewildered, please?

Sorry for jumping in this late, but I've been off enjo^H^H^H^H doing
business at Siggraph....

Seeing as one of my projects is mentioned here, let me toss stuff into
the fire.

"Bullard, Claude L (Len)" wrote:
> It's simple:  comply and compete.  XML has nothing
> in the way of semantics, real time 3D interoperability
> depends on semantics for behavioral fidelity.  

Correct, and that is one of the most annoying things about the X3D
efforts today. Some certain people at trying to turn XML into a
programming language, using attributes as API calls.

> The X3D effort is being revamped.  XML is still there
> as an encoding, but the emphasis appears to be changing.

The reason for this is that XML is incapable of modelling the correct
relationships or data structures. My personal feelings is that XML is
not capable of representing 3D concepts adequately and is a wasted

> the syntax of VRML (what X3D is basically) and more to
> the idea of using the DOM.  This has to be overcome
> with reliable conformant performant implementation.
> So far, no one has stepped up to that except the
> Xj3D group.

As one of the Xj3D people, there is a bloody good reason for not liking
DOM. It never was, and never will be usefull for doing realtime
high-speed work. The event bubbling semantics kills any capability of
dealing with fast changing items, such as time or coordinate
transformations. Another facet of the problem is the text to datatype
and return journey. Cloth and spline animations (eg character animation)
are all but impossible to do with any speed using the DOM view of the
world. Should people be interested, I can expand on the event issues,
but I feel this is not a DOM-oriented list.

> XML doesn't add much to the issues
> of ubiquitous real time 3D at the browser level.
> On the front end (transformation of data sets
> for visualization and high level authoring
> languages) it can be very useful, but at this time
> 3D is mired in the interoperation, fidelity, and
> ubiquity of rendering plugins.

I disagree completely. I think XML sucks as far as 3D representation is
concerned. While I think it works well with 1D and 2D representation, 3+
dimensions it is woefully inadequate for. Dealing with large datasets
(eg sci-viz of hundreds of thousands of data points) and with strict
time synchronisation is a problem (SMIL is less than useful for most
realworld tasks). The strict linear heirarchical model is not capable of
adequately representing the requirements of a 3D rendering system or
dealing with compositional aspects (building and delegating/deriving
scene functionality in an OO-style manner).

Justin Couch                         http://www.vlc.com.au/~justin/
Freelance Java Consultant                  http://www.yumetech.com/
Author, Java 3D FAQ Maintainer                  http://www.j3d.org/
"Humanism is dead. Animals think, feel; so do machines now.
Neither man nor woman is the measure of all things. Every organism
processes data according to its domain, its environment; you, with
all your brains, would be useless in a mouse's universe..."
                                              - Greg Bear, Slant