RE: [xml-dev] Wikipedia on XML

["Jim Tivy" <jimt@bluestream.com>]

In general what I hear on this list is lets not make XML a monolithic beast
that requires a "whale" of a parser.

I think the answer may lie in a pipeline of processors working off a SAX
like emitting parser (not a new idea).

Pipeline nodes:

Node 1 does a content parse - no entity inclusion - expands character
reference.
Other nodes could be:
DTD validators
RelaxNG validator
XSD validator
Namespace annotator

This approach would simplify XML to content only.  Get rid of the Entity
stuff from the core XML specification.

For upward compatibility DOCTYPE would stay as optional (which it is now).
Internal DTDs would not be allowed and DOCTYPE could only reference external
files which would be processed by a DTD validator in the pipeline - if
present - otherwise ignored.

Perhaps we would still get vendors like Xerces putting it all into one
package with flags - but then Apache could release using the a standard
pipeline/event specification.


-----Original Message-----
From: Michael Ludwig [mailto:milu71@gmx.de] 
Sent: Wednesday, August 26, 2009 2:44 PM
To: XML Developers List
Subject: Re: [xml-dev] Wikipedia on XML

Amelia A Lewis schrieb am 25.08.2009 um 20:12:06 (-0400):
> On Wed, 26 Aug 2009 00:03:03 +0200, Michael Ludwig wrote:
> > 
> > So unfortunately, I didn't understand your strategic observations on
> > HTML and XML either. Maybe it's some kind of Bolshevik-Menshevik
> > conflict?
> 
> Oops!  Then please accept my apologies.  Although ... since my
> academic training was in Soviet History, I'm utterly *charmed* that
> you should liken this to the split in the social democratic movement
> in Russia.  :-)

Hmm. To be honest, it wasn't meant to be charming - just an analogy. But
if you find it charming, that's all the better.

> HTML, in all revisions to date, includes a "standard" (actually, a W3C
> Recommendation, which is a nice distinction), and browser extensions.
> XML was designed [...] for extensibility (although 1.0 did *not*
> include the notion of how to distribute authority--that was [...] only
> achieved via the "namespaces in XML" specification [...]

So XML for extensibility, namespaces for compartmentalizing XML work.

> HTML is consequently a bit like freedom of the press

Consequently? But I get your point about "freedom of the press".

While I could do without this "free press", I think that browser vendors
do have stakes in and rights to HTML as they have put a lot of work in
building their browsers and do not charge us a lot of money for it.

> The pressures that influence a person, or corporation, that achieves
> that level of "influence" tends to produce, in those persons, a
> similarity of mindset.

Maybe, but it sounds like speculation to me. If assumed valid, that
statement should be assumed valid for other camps, too.

So MathML, ChemistryML and SVG did not come from the browser camp.

I'm uncertain about the implications of the term "distributed
authority". Aren't namespaces just a technical mechanism for shielding
one collection of words from all others? A practical necessity?

> When authority is distributed, passion and utility become valued of 
> themselves.

Hmm. Aren't they always?

> XML (+namespaces) has that.  HTML does not.  Instead, it has browser
> vendors.

In general, these browsers serve me pretty well. I would also think that
at least some passion has been put in their development.

> It's not so much political, it's just that the HTML specification
> (regardless of version, and excepting XHTML) does not provide a way
> for passionate specification geeks to pound out an interchange format
> in dad's garage (on his automotive computer?) and see it adopted via
> contributed code.  Contributed code *can* do so such specifications in
> an environment that promotes distributed authority.

Would that also promote browser stability? Maybe some of it can be done
via browser extensions?

> > (I always hope my comments trigger insightful replies. Just
> > sometimes, like now, I lack the background necessary to understand.)
> 
> Does this help, I hope?

Yes! I see clearer now. Thanks!

-- 
Michael Ludwig

_______________________________________________________________________

XML-DEV is a publicly archived, unmoderated list hosted by OASIS
to support XML implementation and development. To minimize
spam in the archives, you must subscribe before posting.

[Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
subscribe: xml-dev-subscribe@lists.xml.org
List archive: http://lists.xml.org/archives/xml-dev/
List Guidelines: http://www.oasis-open.org/maillists/guidelines.php