[X3D-Public] Fwd: Re: [X3D] X3D HTML5meetingdiscussions:Declarative 3D interest group at W3C

[Philipp Slusallek]

Hi,

Yes, I am promoting XML3D but I am also defending the general idea to
think about doing 3D the Web way. I am perfectly open to discuss
changing XML3D using better features and approaches -- if people make
concrete suggestions.

And yes, not everything we wish we had is working yet, but this is
expected after only about a year and a half. And given that we have a
reasonable spec, two native (Firefox and Chrome) and one WebGL based
implementation, and a couple of projects that are starting to use it, I
am quite happy with what we have achieved so far. With AnySL and XFlow,
we will soon have a number of capabilities that are not even in X3D yet,
which is not too bad either.

PROTOS and X3D event propagation are not in XML3D, and we are not sure
that we really need the latter, as I have discussed (BTW, at least the
first and I believe also the second is not in X3DOM either, at least
when running integrated with WebGL).

While this probably mean that complex X3D scenes may not be easily
transcribed into running within the Web browser (neither XML3D nor
X3DOM), this backward compatibility is not our main concern, as I
explained earlier. Having said this, it would be nice to be able to
convert as much as possible, though. There may even be way in which at
least some of that functionality can be transcoded, e.g. using some sort
of wrapper scripts for the SAI interfaces mapped to their DOM
counterparts -- but full compatibility seems really hard.

BTW, our next meeting to discuss joint ideas between XML3D and X3DOM is
scheduled for early this year.

	Philipp


Am 02.01.2011 22:15, schrieb GLG:
> Hello Philipp,
> 
> I sense you're a bit exhausted with my "political" arguments
> so I'll try to restrain myself and be more concrete,
> although the more I read and think about your own posts on
> XML3D, the more I feel you are just as heavily promoting
> XML3D. So let's say we're even and get past that.
> 
> Moving on. One of the first things you said was that "making
> the DOM a great declarative 3D scene graph should be our
> main goal here." From this premise, it becomes apparent that
> it is your contention that little if nothing of X3D should
> be salvaged. You do acknowledge that some parts such as
> synchronized events, data propagation and prototypes are
> worth considering under different implementations,
> emulations or simply imitated, but I can't help notice that
> these parts are all but largely missing from XML3D. You
> often point to incomplete or planned work, or to obscure
> other items that are also not quite ready but expected to
> work. Plus, we have barely touched the subjects of how
> complexe, multi-layered and flexible PROTOS can be, and how
> a simlulation would work across domains. In short, a lot of
> conjectures, expectations and assumptions, to pretty much
> start from scratch on the way to a 3D web. Not only would it
> be necessary to assemble, debug, rebuild, upgrade a number
> of parts to make everything work, we also have to contend
> with the fact that there is no upgrade path in sight nor
> even a hint of compatibility with existing standards. We can
> transcode 3D objects but most if not all existing
> interaction and behavior would be lost. In essence, this is
> like going back to VRML1 without any real reason to believe
> this would result in a wider acceptance of 3D on the web
> (Let's not forget that VRML1 and VRML2 were very simple to
> use). Isn't that a whole lot of wishful thinking? 
> 
> On the other hand, X3DOM based on WebGL is also being
> proposed. WebGL which is really like an evolved subset of
> OpenGL, the later which has always been the underlying
> foundation of VRML and X3D (I don't have to tell you this
> and I'm simplifying but not much). X3DOM itself is
> recognizable X3D. So nothing is radically different here,
> and existing artwork have a better chance of salvation. It
> also does achieve the goal of rendering 3D in a web browser,
> with the distinct advantage that an upgrade path to full
> fledged X3D will remain available for as long as we need and
> want it.
> 
> We are forced to generalize for the sake of brevity, but
> what am I missing here Philipp. I find it extremely
> difficult to get my head around to accepting XML3D. That is
> without mentioning the potential loss of 14+ years of
> development in interactive VRML/X3D. Not just my work, but
> countless others. It is a very big step and a huge leap of
> faith you are asking. And for what? Really the benefits are
> not all that clear, and performance would suffer
> tremendously IMO. I find it hard to imagine running a world
> like Office Towers over Javascript and CSS in a web browser.
> That just seems ludicrous. There has to be a better
> solution. I really think 3D in a browser should be a gateway
> step to more solid applications when it's time to run
> processing intensive, memory gobbling worlds; at least for
> some time until it all get sorted out.
> 
> I hope these were technical issues enough. Please do not
> give up on me just yet. I am a reasonable person. As long I
> am still listening, I can be convinced if I believe the
> arguments presented. Perhaps you could talk about what parts
> of XML3D you would be willing to remove in favor of X3DOM,
> so that more people would like the outcome.
> 
> Lauren
> 
>