W3C home > Mailing lists > Public > www-tag@w3.org > April 2003

Re: XPointer [was: First Draft of summary of TAG issue abstractComponentRefs-37]

From: Chris Lilley <chris@w3.org>
Date: Thu, 10 Apr 2003 18:13:36 +0200
Message-ID: <108168171656.20030410181336@w3.org>
To: Paul Grosso <pgrosso@arbortext.com>
CC: www-tag@w3.org

On Thursday, April 10, 2003, 4:15:43 PM, Paul wrote:


PG> At 21:59 2003 04 09 -0700, dorchard@bea wrote:

>>4. Use full XPointer.  The sample URI is
>>http://airline.wsdl/ticketagent/#xmlns((w=http://schemas.xmlsoap.org/wsdl/)x
>>pointer(//w:portType[@name="TicketAgent"]/w:operation[@name="listFlights"]/w
>>:input[@name="listFlightsRequest"])
>>
>>Pros:
>>- re-use XPointer syntax, which is a rec

PG> No, it [1] is not.

These parts of XPointer are Recs

25 March 2003: XPointer element() Scheme
http://www.w3.org/TR/xptr-element/
25 March 2003: XPointer Framework
http://www.w3.org/TR/xptr-framework/
25 March 2003: XPointer xmlns() Scheme
http://www.w3.org/TR/xptr-xmlns/

which may be the source of the confusion.

So you can point to elements, you can point to elements in a given
namespace, and you have a framework for developing new schemes, as
Recs.

The full, Xpath-like path with DOM2-like range, xpointer scheme is still
a working draft.

Unfortunately yes, that means you cannot point at particular attributes
or point at particular elements based on them having those attributes.
Which would seem the next step up in complexity from element(),
without going all the way to range. Indeed, you cannot even point to
an ID except using the bare name form.

This also implies that SVG should deprecate the non-barename way of
pointing to IDs, which we speculatively adopted as a way of saying 'we
are not assuming full xpointer in this version, but we might well once
it is done'.

OTOH it does mean that the XPointerIDRef has been a Rec since
September 2001 ;-)
http://www.w3.org/TR/2001/REC-SVG-20010904/linking.html#SVGFragmentIdentifiers
although it points to th Xpointer spec.

Question is should we point to the XPointer scheme spec for this? Or
should we make an svgid() scheme? Or should we make an xmlid() scheme?
Or should barename references justbe the way to go?

Not that this helps when referring to other attributes, as Dave's
examples do.


PG> It is not even at Last Call yet, and there is no currently
PG> existing working group responsible for working on it.

Agreed.

PG> [1] http://www.w3.org/TR/2002/WD-xptr-xpointer-20021219/



-- 
 Chris                            mailto:chris@w3.org
Received on Thursday, 10 April 2003 12:13:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:17 GMT