groves dissent (was RE: RFC: Attributes and XML-RPC)

[Paul Prescod]

Reynolds, Gregg wrote:
> 
> I have a few basic problems with it (leaving aside the
> question of English prose - don't get me started).  First, it forces
> everything into a two-dimensional tree.  

In what sense is it more two-dimensional than a DOM? I guess I'm not
even sure in what sense it is "two dimensional." I can depict a grove as
a stream of bits (one dimensional) or as a flat tree (two dimensional)
or a 3D tree or...okay, I have trouble going beyond that, I admit.

> Then, since that doesn't really
> work very well, given that attributes and children (in the ordinary,
> intuitively understandable sense of these terms) are different types of
> critter, it distorts this by privileging a certain "property" (read
> "child"), and thereby a particular set of paths in that tree, as the real
> tree.

Other ways to think of this are as a distinction between:

content/metadata
ownership/containment

You own your attributes but you contain your sub-elements. I don't see
this as very complicated. It allows every object in the grove to be
enumerated in a straightforward manner. The DOM has a similar concept:
the content property is always called childNodes. Note that even for
XML, the grove is more consistent with the XML standard's terminology.
The "content property" of an element is just called "content" (as in the
XML specification).

> In a word, it takes what is basically pretty simple - an attributed
> tree - and turns it into something of surpassing obscurity. 

But most XML documents are *not* just attributed trees. How do you
handle the doctype? How do you handle ID/IDREF? How do you handle
inter-document hyperlinks? The DOM doesn't answer most of those
questions yet. (am I demented or is there REALLY no efficient way in DOM
1.0 to dereference an IDREF??? wait...don't answer that...just tell me
if there is a way to dereference an IDREF).

> But "the grove
> model" isn't even necessary.  The thing it tries to model can be modeled
> quite adequately without any invented terms or concepts.

I think the comment deserves more explanation. Where can we get a data
model for hyperlinked media without inventing anything? (er...ISO!)

> But not all great experiments succeed, and groves/DSSSL/Hytime have
> failed in the marketplace for good reasons, and that should guide us in
> building XML.

It is quite frightening, the speed with which things can be declared
"dead" these days. It took object oriented programming twenty years to
catch on. Pretty soon the mainstream press will declare XML dead.

 Paul Prescod



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)