XSchema Spec, Section 3, Draft 1 (Namespaces)

[Peter Murray-Rust]

At 22:11 30/06/98 -2800, John Cowan wrote:
>james anderson scripsit:
>
>> as you've said, i'm repeating myself too much, and without effect.
>> i'll shut up about this 'till i figure out how to make it clear.
>
>You *are* clear.  However, you want the namespace-aware parser to
>do something that namespace-aware parsers aren't contracted to do:
>decide what is, and what is not, an "ultimate" reference to a
>universal-name, and translate all such.  The authors of the namespace
>draft are a proper subject of this complaint.
>
I sympathise with those trying to define how namespaces are used under
XSchema and elsewhere. Any XML-SIG discussions are confidential, and I
don't have a timescale for the next draft release. However I think it can
be reasonably said:
	- there are many different views on what namespaces are and how valuable
they are.
	- the current approach is deliberately minimalist
	
The WhiteKnight question is always with is:
	- what is a namespace?
	- what is a namespace called?
	- what is the name of the namespace?
	- what is the name of the namespace called?

This calls for great precision and agreement which I suspect we do not yet
have. (It also arises in the PubID vs SysID question). 

	I suspect there are two strategies:
		- accept it as a very difficult problem and not attempt to solve it. The
community (market) will then come up with approaches which may or may not
find ecological niches
		- accept it as a very difficult problem and propose a solution. Some of
the community will accept this and some won't.

I would suggest that we only address the mechanism whereby prefix is linked
to ns. And that in a minimalist manner (i.e. providing Names for these
beasties). *What* ns should refer to is beyond XML-DEV at present (though
it may be a very appropriate forum in the future).

FWIW I will adopt a very simple strategy for JUMBO. I will not use ns to do
other than to give me a unique string, thus:

<?xml:namespace ns="xml-cml.org" prefix="CML" ?>

The ns value is unique within the galaxy. I then resolve it with a
JUMBO-specific PI:

<?JUMBO ns="xml-cml.org" java="jumbo.cml"?>

This will *implicitly* link any CML element to a class:

<CML:MOL> links to jumbo.cml.MOLNode

(JUMBO will have the convention that <FOO> links to FOONode). These classes
will be stored under the classpath (using the tools that you have all
helpfully told me about recently).

I will also store the CML XSchema in the same area. Whether XSchema need to
give me a handle for this I don't know (I shan't use 'src'). *** This
brings up the question of what the XSchema file is called ***. Do we name
it after the 'root' element? Thus:
	- what is the XSchema for CML called (a) if there is a <!ELEMENT CML ...>
(b) if there isn't?
	- what mechanism might there be for locating the XSchema file?

	P.

Peter Murray-Rust, Director Virtual School of Molecular Sciences, domestic
net connection
VSMS http://www.nottingham.ac.uk/vsms, Virtual Hyperglossary
http://www.venus.co.uk/vhg

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/
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)