W3C home > Mailing lists > Public > www-tag@w3.org > June 2007

Moving xmlFunctions-34 forward

From: Henry S. Thompson <ht@inf.ed.ac.uk>
Date: Mon, 11 Jun 2007 16:13:34 +0100
To: www-tag@w3.org
Message-ID: <f5by7iq34o1.fsf@hildegard.inf.ed.ac.uk>

Hash: SHA1

Wrt " The Interpretation of XML documents " [1] (TimBL, 2005-08-27)
and the discussion at the TAG f2f in February [2]:

TimBL somewhat mis-represents the 'pipeline model' -- it's primarly
about doing _externally_ specified computations over XML documents,
_not_ about interpreting/implementing their intrinsic semantics.

The confusion arises because with the advent of things such as
XInclude and XML Encryption, the possibility arose of _using_
pipelines to 'get at' the 'real' XML document from which
interpretation should start, i.e. by doing XInclude processing and
decryption.  TimBL is right to point out that there are problems even
here, to do with ordering and the unavailability of (repeat until
done). . .

XSLT LRES, and anything which uses an attribute as a signal, has a
problem because there's no way to tell what is nested in what.
TimBL's preferred alternative reads to me like fx:transform. . .

*Top-down Processing model* -- 2nd bullet is what elaboratingInfosets
tries to define.  Deferred processing is correctly identified as an
optimisation issue, not a definitional one.  eI specifies what the eI
of a document _is_, but doesn't require you to compute it all at
once.  The requirement that eNSes are compositional allows for this.

Quoting and control:  The discussion in Cambridge in March foundered
on the issue of the relationship between the elaboration story as I
wrote it and the promise of a full compositional/recursive-descent
story about XML document semantics in general.

In particular, I take it, TimBL would like a particular XML vocabulary
to make its own decisions about when to use the elaborated infoset and
when, effectively, to quote a subtree.

So, would the following be more likely to attract consensus:

 Change the story in the draft to include a notion of 'constrained
 elaboration', during which a specified set of namespaces act as
 quotation signals.

That would allow any spec., associated with one or more namespaces, to
completely control when elaboration was and was not performed.

GRDDL or W3C XML Schema might use full unconstrained elaboration,
whereas RDF might want to identify the RDF and OWL namespaces as
quotation signals, and choose which non-RDF/OWL subtrees to elaborate
and which to leave 'as is'.


[1] http://www.w3.org/DesignIssues/XML.html
[2] http://www.w3.org/2001/tag/2007/03/06-minutes#item10
- -- 
 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]
Version: GnuPG v1.2.6 (GNU/Linux)

Received on Monday, 11 June 2007 15:13:36 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:52 UTC