Re: [xml-dev] XML Redux

[Liam R E Quin <liam@w3.org>]

On Wed, 2011-02-16 at 20:17 +0100, Danny Ayers wrote:

> XML itself isn't redundant, far from it, there are plenty of things
> like Atom, DocBook, ODF etc for which it seems a good fit. But in
> cases like that, would a simplified XML work as well?
> 
> I guess I don't really see what role a simplified XML could play that
> isn't already adequately covered.

There are a lot of "gray areas" between document and data where a
mixture is a win. Remember that XML excels at mixed content, so that it
can represent things like titles of mathematical books and papers, even
if only as an embedded string in RDF or (as Mike Kay proposed) JSON.

Atom uses a tiny fraction of XML functionality (when was the last time
you saw an RSS feed with an NDATA entity in it?) and so do most other
formats, and, frankly, it's pretty much the same subset.

The big challenge is not losing the interoperability -- unlike when we
made SML from SGML, there are now billions (literally) of deployed XML
parsers in the wild, using APIs that are not easy to change.

A new API that's, say, more efficient and easier than DOM could take off
fast (e.g. see the success of jQuery: it made the programmer's life
easier) but the old APIs will still be there.  A simplified subset that
is a true subset has some attraction, because the newer APIs can be
simpler too.

See also my XML Redux blog posting for some background to this thread.

Liam

-- 
Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/
Pictures from old books: http://fromoldbooks.org/
blog - http://www.barefootliam.org/