RE: [xml-dev] Towards XML 2.0

["David Lee" <dlee@calldei.com>]

"Pretty" isn't necessarily the point.
By defining an absolutely bare bones processor we can have "XML" implemented
extremely efficiently, and remove lots of variation.
I agree I'd rather TYPE attributes then elements, and certainly not all
classes of documents would benefit from these.
But by say removing attribute support *in a particular profile* both the
processor AND the data generation could be vastly simplified.
And sure maybe XPath doesn't even *work* in such a simplified "XML".

But as  long as the data is parable by a 'more complex' profile then I would
still call it XML.

e.g.
	<foo>bar</foo>

I call "XML"  even though it doesn't contain any attributes, entities,
namespaces, comments, notations or mixed content.
And its parable by even the most fully blown processor.

IMHO it could be *vastly useful* to have these micro profiles.
The intent isn't to please everyone within *each profile* but rather to
define profiles that *please a certain class of producers and consumers* and
still be within the "XML World".

If you don't like that profile don't use it.   But if we define some of
these 'very basic' profiles we may well see willingness to embrace "XML" in
cases they are not today, such as places JSON is used and need not be.
Mobile devices.  Embedded devices.   Web pages that don't take forever to
load/run.
Apps of all sorts that would like "some features of XML" but don't want to
buy into the whole enchilada when they need only a piece of cheese.

Imagine the size and speed of a processor if it only had to handle a few
things instead of even the "Minimal" profile proposed by the Profile Spec ?
It could possibly be as little as a dozen lines of JS code (? just guessing
?) and still be "conformant" to that profile.

Tool Vendors could provide a range of processors with increasing profile
support (and size and slowness) and you could pick the one you *absolutely
needed* for the particular environment and not be required to lug around the
elephant "just because it's XML".... 




----------------------------------------
David A. Lee
dlee@calldei.com
http://www.xmlsh.org


-----Original Message-----
From: Andrew Welch [mailto:andrew.j.welch@gmail.com] 
Sent: Tuesday, December 07, 2010 10:05 AM
To: David Lee
Cc: Pete Cordell; vojtech.toman@emc.com; xml-dev@lists.xml.org
Subject: Re: [xml-dev] Towards XML 2.0


On 7 December 2010 14:42, David Lee <dlee@calldei.com> wrote:
> I'd even argue for a minimum *without attributes* and without mixed
content,
> no DTD subset, no namespaces, and only UTF8 support.

I've had to deal with "element only" xml and it's not that pleasant...
instead of the usual:

<item type="foo">value<item>

you have the verbose:

<item>
  <type>foo</type>
  <value>value</value>
</item>

and your xpaths go from:

item[@type = 'foo']

to

item[type = 'foo']/value

...so not great.

I agree with no DTDs, no namespaces, UTF-8 only etc... mixed content
has to stay otherwise you may as well not call it xml.


-- 
Andrew Welch
http://andrewjwelch.com