[X3D-Public] Fwd: Re: [X3D]X3DHTML5meetingdiscussions:Declarative 3D interest group at W3C

Len Bullard cbullard at hiwaay.net
Sat Jan 1 11:23:59 PST 2011


Then the proof is the implementation.  I didn't say, unuseful, but
pragmatically, the stylistic uses of 3D in HTML have been limited where
available.  Here I am not referring to X3DOM which is useful.  It works.

All renderings whether plugin or hanging on the DOM have a richness
quotient, that is, what can be usefully done with it.  I am a world builder
and it's an old and rich form but not the only useful form.   And that is a
question the marketing people at least should be asking: what forms are
effectively displayed in which engines not lingua because frankly, that's
all syntax in XML and in any application of XML, local from XML's
perspective.

So the event model and the DAG vs. tree are points of contention.

The tree vs DAG is a problem of attribute parsimony.  An XML att is a list
and structural information inside the att list is largely local
(microparsing aka make up your own rules for this string).  There is a good
case for attribute structures to become rich.    This is a problem for XML
not just from X3D but realistically, object oriented languages. That's an
XML dilemma.  The W3C's problem it is.

The event model is the challenge.   In short and possibly naïve terms, the
key of the current DOM API's I've used for X3D (and you're the expert), it
comes down to DOM-to-timers.  The event cascade is handled by the X3D
renderer.  That's easy but then come the rules for cycles in the DAG linked
object for which one implementation scripts those directly to variables in
the code and because of VRML legacy, X3D has ROUTEs but the better
implementations support both.  Will HTML3D handle that in CSS?

Do ROUTEs survive a merging?  They are convenient but hazardous in
authoring.  It is a major hassle to be working in the script and editing
variables that invalidate ROUTEs and not having the variable method is the
comfortable linguistic scripting form.   The trouble of ditching ROUTEs is
they are a dip-simple means to declare input/output connectors among the
language local forms, ie, the graphic objects, so in that linguistic domain,
convenience in that the scripts that support them are not exposed to the
author or editor.  Because behavioral scripting is the most technically
difficult part of the practice, having a means to hide that is useful.  On
the other hand, common practice is to put that information in the GUI
properties. 

CSS ROUTEs?

len

-----Original Message-----
From: Joe D Williams [mailto:joedwil at earthlink.net] 

> HTML doesn't make
> X3D better.  It makes it stupider.

I guess we for sure have to watch out for that. To me the big item 
here would be an XML form that can be schemafied and proven, easy 
transcoding from whatever form to X3D and vice-versa where that 
commonality of concepts and data applies. I recall that we did not 
really have an XML encoding for X3D until we showed typical XML 
transformation techniques could be used to go back and forth between 
VRML2 and XML X3D and until we stretched the XML schema to get the 
types of data we needed.

Thanks and Best Regards,
Joe




----- Original Message ----- 
From: "Len Bullard" <cbullard at hiwaay.net>
To: <info at 3dnetproductions.com>; <x3d-public at web3d.org>
Sent: Thursday, December 30, 2010 10:33 PM
Subject: Re: [X3D-Public] Fwd: Re: 
[X3D]X3DHTML5meetingdiscussions:Declarative 3D interest group at W3C


> They don't bother me although I understand you position.  It's just 
> that
> like clockwork someone from a consortium or faux with bigco pockets 
> has been
> coming round here for a decade and a half telling us they can do 
> this better
> and so far, none have.  It's getting silly.
>
> For X3D, it's about content now.  Systems work.  HTML isn't a good 
> host and
> XML is too agnostic.  The original idea of a solid DAG engine with a
> cascading event model and internal property sets works better for a 
> scalable
> set of applications than anything I've seen proposed.   HTML doesn't 
> make
> X3D better.  It makes it stupider.
>
> len
>
>
> From: x3d-public-bounces at web3d.org 
> [mailto:x3d-public-bounces at web3d.org] On
> Behalf Of GLG
>
> I know. I'm trying to be consilient so we avoid getting into
> another blood bath. Lauren
>
>
>
> _______________________________________________
> X3D-Public mailing list
> X3D-Public at web3d.org
> http://web3d.org/mailman/listinfo/x3d-public_web3d.org 




More information about the X3D-Public mailing list