W3C home > Mailing lists > Public > www-ql@w3.org > October to December 2003

Re: namespace node implementation

From: james anderson <james.anderson@setf.de>
Date: Fri, 24 Oct 2003 21:49:56 +0200
To: www-ql@w3.org
Message-Id: <3E0DFAD5-065B-11D8-96EF-000393BB8814@setf.de>


On Friday, Oct 24, 2003, at 18:38 Europe/Berlin, Kay, Michael wrote:

> I thought this discussion was about how to implement the current 
> specifications concerning namespace nodes.

it was. in particular, it concerned whether proposed changes foster 
implementation or impede it.

>  You seem to have changed the subject, and to be talking about 
> alternatives to the current specification, and indeed to be engaging 
> in philosophical discussions about the merits of namespaces in the XML 
> world view.

there was no subject change. perhaps the relevance was not obvious. if 
the "disaster" citation or the distinction between "truth" and 
"misfortune" were distracting, feel free to delete them. they have no 
bearing on the substantive issues. the remaining text discussed 
concrete implications of features of proposed specifications. there 
were several observations.
1. the proposed models are not inherent in xml decoding/encoding 
requirements.
2. the proposed models fail under combination operations.
3. namespaces 1.1 will hamper the use of dtd with namespace aware 
documents and codify their obsolesence.
4. simpler models satisfy the decoding/encoding/processing requirements 
and do not suffer from modeling problems.

>
> I'm afraid I can't cope with all three levels at once. You'll need to 
> be clearer about which plane we're on.

i'm a bit surprised at a tack in the direction, "the question was too 
amorphous to cope with." the thought problem was basic, was not at all 
philosophical, and is an essential discriminant between a model which 
is an adequate basis for xpath/xquery and one which is not. given the 
number of planes which are inherent in "namespace aware xml", the 
thought question was pretty elementary. if there were too many levels, 
then perhaps it is time to take a step back, admit that there are too 
many levels inherent in the current approach to namespaces, and redraft 
namespaces 1.1 to specify a 1-1 relation between prefixes and iri.

...
Received on Friday, 24 October 2003 15:50:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:17:16 UTC