xml-dev - Re: The Power of Groves: The VRML View

[Len Bullard <cbullard@hiwaay.net>]

Steve Schafer wrote:
> 
> What I'd like to see is a reasonably concise description of how XML
> was found to be unsuited to the task. 

What I would like to do is use the grove modeling first.  We know 
XML.  Groves are the obscure model yet they are claimed to have 
great representational power.  The challenge is to apply them 
and see.

> What specific features (or lack
> thereof) of XML notation made it inappropriate for VRML? What kinds of
> structural relationships were you trying to model for which XML
> notation was too cumbersome?

The notes at the clue.  It isn't that XML is too cummbersome. It 
is the mapping of the meta-structures.  Hint:  it is obvious to 
map XML elements to VRML nodes and XML attributes to VRML fields. 
If you try that, how much and what kind of information do you lose 
if you do?

> Where I'm coming from:
> 
> Unlike Peter and probably most of the others participating in this
> discussion, I'm actually very comfortable discussing all of this in
> abstract terms. (Perhaps it's because I'm a physicist by training--I
> don't know.) 

If you don't know you may be less trained or more uncomfortable than you 
think.  Regardless, the point of this is to bring everyone along with 
the discussion, to make all of the thread participants comfortable 
with groves in order to evaluate their suitability for untangling 
the W3C language specifications and any others.  

> And I do strongly feel that looking at the abstract
> picture is the way to understand precisely where any deficiencies may
> lie.

Not necessarily first.  MMTT disease sets in quickly.
 
> I don't think we can build
> usable tools unless we have a firm, formal foundation, one which will
> inevitably include numerous abstract models. Without that foundation,
> we're just chasing our tails. 

I agree.  Eliot agrees.  All of the original community that created 
groves, HyTime, DSSSL and SGML probably agree.  That is why they 
designed groves.  

> We need to make this stuff accessbile, but first we need to make it
> work. 

Chicken and egg and precisely what the designers of HTML told the 
members of comp-text-sgml.  HTML works.  So does HyTime.  You see, 
we can go in either direction.  

> I'm an experienced software developer; I've written parsers,
> interpreters, sophisticated text and graphics rendering engines, etc.
> Complex modeling and programming issues don't scare me. What scares me
> is the possibility that I'm going to embark on a major journey using
> groves as a fundamental data abstraction for a very large project, and
> then a year from now I'm going to hit a brick wall because of some
> unforeseen deficiency.

Then I would say the best thing is to help those of us who are not 
so well trained to understand this so we will not as a mass embrace 
something that will trouble you or us later.  I am sorry if that is 
not a sophisticated task, but it is a necessary one.  If I understand 
what Murray, Andrew and others are saying, they are about to go 
full speed ahead with X-schemas with the same marketing power and 
relish as they went full speed ahead with HTML.  

Yet we had to create XML.

So for some short time, it doesn't seem that taking groves by 
pertinent examples from atoms to complex structures step by 
step is too much to do.  It may be spilt milk, but reinventing 
the wheel every three years is expensive.

len