W3C home > Mailing lists > Public > public-qt-comments@w3.org > November 2004

RE: [DM]Relative order of Attribute Nodes

From: Mukul Gandhi <mukul_gandhi@yahoo.com>
Date: Thu, 4 Nov 2004 09:41:20 -0800 (PST)
Message-ID: <20041104174120.65284.qmail@web41307.mail.yahoo.com>
To: Michael Kay <mhk@mhk.me.uk>, 'Michael Rys' <mrys@microsoft.com>, public-qt-comments@w3.org

Thanks for insights! So it will mean that, the
expression @*[n] is not much useful.. As it might
produce unpredictable results.. My desire to have the
output of @*[n] predictable, is due to the intutive
nature of the expression! 

One of my learning from this thread is that, not to
develop XSLT applications that rely on original
lexical order of the attributes..   

Regards,
Mukul

--- Michael Kay <mhk@mhk.me.uk> wrote:

> If you (the stylesheet author) want the attributes
> in a particular order,
> you can sort them, e.g. by name. 
> 
> You can't reproduce the original lexical order of
> the attributes because the
> information is not available from the XML parser
> (more formally, from the
> InfoSet).
> 
> Michael Kay
> 
> > -----Original Message-----
> > From: Mukul Gandhi [mailto:mukul_gandhi@yahoo.com]
> 
> > Sent: 04 November 2004 14:18
> > To: Michael Kay; 'Michael Rys';
> public-qt-comments@w3.org
> > Subject: RE: [DM]Relative order of Attribute Nodes
> > 
> > I agree with you.. From XML's point of view it
> > certainly makes sense (as you have explained).
> > 
> > But I still do have a feeling(from XSLT point of
> view,
> > and not XML), that making output of @*[n]
> "consistent
> > across XSLT implementations" would be useful for
> some
> > classes of problems.. i.e. @*[1] should return the
> 1st
> > attribute, @*[2] the 2nd one and so on..
> > 
> > Is it worth debating about this feature, and if
> found
> > useful be made part of the "XQuery 1.0 and XPath
> 2.0
> > Data Model"..  
> > 
> > Regards,
> > Mukul
> > 
> > --- Michael Kay <mhk@mhk.me.uk> wrote:
> > 
> > > 
> > > > Thanks for the answer. I am now curious, why
> XML
> > > > attributes have this characteristic..
> > > 
> > > That's a historical question: you might get a
> better
> > > answer from the SGML
> > > folks on xml-dev. There is, I think, a very
> strong
> > > consensus in the XML
> > > world that elements are ordered but attributes
> are
> > > not - though I don't
> > > think there is anything in the XML spec itself
> that
> > > says so. However, you
> > > will find this consensus reflected in most
> > > processing models and APIs for
> > > XML. It derives, I think, from the traditional
> use
> > > of elements to represent
> > > the visible text in a document and attributes to
> > > represent its typographical
> > > properties.
> > > 
> > > Michael Kay



		
__________________________________ 
Do you Yahoo!? 
Check out the new Yahoo! Front Page. 
www.yahoo.com 
 
Received on Thursday, 4 November 2004 17:41:52 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:57:02 UTC