RE: Options for dealing with IDs

I agree we are talking about things on different planes.  I care about how
an actual parser would work, not just how specs work.  And I believe it's
important to consider the impact on building parsers.  I consider the
infoset a glossary, not an encapsulation mechanism/magic layer that means we
can or should ignore how the parser is built.


> -----Original Message-----
> From:
> []On Behalf Of
> Paul Grosso
> Sent: Tuesday, January 07, 2003 2:28 PM
> To:
> Subject: RE: Options for dealing with IDs
> At 14:18 2003 01 07 -0800, David Orchard wrote:
> >Sure XPointer looks at the infoset.  But how does the
> infoset get the idattr
> >property?
> Maybe we are discussing things on different planes.
> You appear to be talking about writing code to process
> XPointers by looking at raw XML.
> I'm talking more about the specs.
> My point is that the XPointer specs are already written
> based on the infoset, and they (as specs) don't need to
> care how things got into the infoset.
> > If I build a streaming parser for XPointer that uses xml:idattr,
> If you build something that isn't an XML parser, you're on your own.
> I see no reason for W3C specs to care about that scenario.
> If you build an XML parser and it doesn't provide the infoset with
> the necessary values, then that parser cannot be used in applications
> that require such values.  The XPointer specs right now make clear
> what values they require from the infoset.  A processor that doesn't
> provide the necessary infoset is not a compliant XPointer processor.
> paul

Received on Tuesday, 7 January 2003 17:48:43 UTC