W3C home > Mailing lists > Public > public-xpointer-registry@w3.org > October to December 2005

RE: xpath/xpath1

From: Michael Kay <mike@saxonica.com>
Date: Sat, 10 Dec 2005 18:55:09 -0000
To: "'Henry S. Thompson'" <ht@inf.ed.ac.uk>, "'Simon St.Laurent'" <simonstl@simonstl.com>, "'Sid Bauman'" <Syd_Bauman@brown.edu>
Cc: <public-xpointer-registry@w3.org>
Message-ID: <E1El9sS-0003H0-EK@lisa.w3.org>

I would have thought it was sensible to define two schemes: xpath1() and
xpath2().

The extent to which the two are compatible depends rather on what one does
about typed data. It would seem desirable that the meaning of an XPointer
should not change depending whether or not the document has been validated
against a schema; this would imply defining the xpath2() scheme to work on
the data model obtained from the document's Infoset, not from its PSVI. In
this situation the vast majority of path expressions would work identically
in both schemes. However, there would be a few differences: for example
within a predicate, the xpath2() scheme would interpret [@a > @b] as a
string comparison whereas xpath1() would interpret it as a numeric
comparison. 

In answer to Syd's question:

>It will be misleading to have xpath1() if there never is an XPath 2,
>3, or 4. If W3C is confident that XPath 2 will become a full-fledged
>recommendation, then I think I'd even prefer to go with xpath1().

XPath 2.0 is now a Candidate Recommendation. One should almost put the
question the other way around: why not just define an xpath2() scheme, and
treat xpath1 as history?

Michael Kay
http://www.saxonica.com/

> -----Original Message-----
> From: Henry S. Thompson [mailto:ht@inf.ed.ac.uk] 
> Sent: 09 December 2005 14:08
> To: Simon St.Laurent; Sid Bauman
> Cc: public-xpointer-registry@w3.org; Michael Kay
> Subject: Re: xpath/xpath1
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> What shall we do about xpath and xpath1?
> 
> Mike, to what extent is xpath2 a superset of xpath1, in terms of the
> node set selected?
> 
> I'm wondering whether using 'xpath' as a scheme name for only XPath
> v. 1 paths will be seen to be downright wrong in due course, or just
> misleading. . .
> 
> A simple example of a path/document pair where the
> XPath-1.0-selected-nodeset is easily seen to be different from the
> XPath-2.0-seelcted-nodeset would be a pretty clear 
> counter-indication. . .
> 
> ht
> - -- 
>  Henry S. Thompson, HCRC Language Technology Group, 
> University of Edinburgh
>                      Half-time member of W3C Team
>     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 
> 131 650-4440
>             Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
>                    URL: http://www.ltg.ed.ac.uk/~ht/
> [mail really from me _always_ has this .sig -- mail without 
> it is forged spam]
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.6 (GNU/Linux)
> 
> iD8DBQFDmY/OkjnJixAXWBoRAhMIAJ9gOJk9LpEN3mVLMdb4y/zyqRlxIwCffI5e
> YWb5OredFCEwVdgFaDlVVKs=
> =gavr
> -----END PGP SIGNATURE-----
> 
Received on Saturday, 10 December 2005 18:55:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:22:06 GMT