Lists Home |
Date Index |
On Tue, 2002-03-05 at 02:53, Eric van der Vlist wrote:
> Title says it all, the extensibility of XML is one of its myths...
I'm not sure it had to be mythical, nor am I convinced that
extensibility is lost to all of use at this point.
> Technically, XML is based on trees which are not the most extensible
> structures (compared to tables or triples). If you extend a tree you are
> likely to break its structure (and existing applications). I would say
> that trees grow but are not "extended".
For some applications, you may be right, and different approaches like
RDF probably make sense in those cases. Still, I have a very hard time
imagining writing my data in triples, and a harder time imagining
triples having anywhere near the transparency that relatively flat tree
Extending trees isn't difficult so long as you haven't tightly bound
yourself to a particular vision of how the tree must absolutely
postively precisely be structured. If you're willing to accept that
adding a new branch to a tree or reorganizing a branch doesn't
automatically make it a diabolical mutant, there's a lot more
Extensibility scares people - while object-oriented programming, for
instance, permits all kind of extensibility through class and interface
structures, it also has best practices which strongly encourage
encapsulation, hiding all of those details from other parts of the
program. XML has a tougher challenge here, as it is both extensible and
I prefer to think of XML as a broad syntax containing all of its
possible applications rather than as a toolkit for building specific
vocabularies. This perspective seems to encourage rather different best
practices than focusing on vocabulary-building. The document contents
are the foundation of processing, not a particular schema, and the same
contents need to be visible ("fully composed", as Tim Bray put it
yesterday) to all comers, without any vocabulary-specific knowledge.
Once we've got that document, we can (and should) go anywhere. Tools
like schemas, stylesheets, RDDL, and code should let us explore and
describe possibilities, not choke off anything that doesn't conform to a
> This technical limitation has been strengthen by a community which has
> become conservative and any evolution seems deamed to fail.
I think the conservatism has always been there. As radical as XML is
("Hey folks! Create your own labeled structures for information!"), a
lot of people have sold it short. Incessant talk about the need for firm
contracts between parties, a fondness for expert committees, and the
continuing desire to couple program logic and information as tightly as
possible are constant themes.
To some extent, it's reasonable. Hurling XML into the world without
such restraints would likely have created (even more of) a backlash
against these crazy anarchists. Programmers are pretty conservative
when it comes to data structures, and I suspect most of them have
learned firm lessons from the limitations of earlier systems. Tools
vendors are out to make money selling tools, not extensibility.
I'm can't say I'm very impressed with the "official" line of evolution
right now. I suspect the W3C is (as it seems to have always been) mired
in a notion of XML as vocabulary toolkit rather than a syntactical
continuum, and their output seems to confirm that. Recent discussions
on www-tag about how tightly to bind processing resources to namespace
URIs aren't exactly encouraging, either.
(On the other hand, I'll admit to liking some of the vocabularies
they're building there.)
> XML is now legacy. Its users community is screaming against any change
> and its specification body seems paralysed by its structure and the
> diverging interests of its members...
I'm not sure that makes XML as a whole legacy. It certainly means that
those of us who want to do things differently have to get our points
across in both words and code. I don't see the W3C or the tools vendors
doing it, and I don't see the people for whom XML is just a small part
of their job doing it.
> It's probably time to look for the next wave!
I suspect the next wave will still be markup of some kind. There may be
tuples or triples involved somewhere, but I can't see them being on the
surface. It may be a good time to start making waves, while the vendors
are all in a corner marveling at Web Services.
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!