New draft of Elaborated Infosets document (xmlFunctions-34)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Further to discussion at the March 2007 f2f [1]. I've produced a new
draft of _The elaborated infoset: A proposal_ [2].

As several editorial notes therein suggest, I'm not at all sure how to
integrate this into Web Architecture.  There are at least two
(partially interdependent) questions this draft doesn't answer:

1) What is the role of the application in controlling the elaboration
   process?

2) What is the right model for the relationship between elaboration,
   understood as the specification/construction of an infoset, and
   application semantics, particularly in the case of mixed namespace
   documents?

To illustrate the reason (1) is important, consider a bank transfer
document, expressed in XML, and three XML applications:

 a) An XML editor;
 b) An XML validator (pick your favourite schema language);
 c) An XML banking application.

The editor probably wants the document _as is_, without  any
elaboration.  The validator probably wants as much elaboration as
possible (but really that's up to the user).  The banking application
may want some but not all of the document to be elaborated:  consider
the case where its own banking markup semantics includes some form of
encapsulation.  So the document can't be said to have _an_ elaborated
infoset:  elaboration is in part controlled/parameterised by what the
user of the document is _doing_ with it.

The issue behind (2) is discussed at some length in Tim Berners-Lee's
original thinking on this topic, as expressed in [3], particularly the
sections entitled "Top-down Processing model" and "Software designs
for top-down processing".  The granularity of the implied interaction
in that discussion is much finer than that in the draft elaboration
proposal.  It's not clear to me that a notion of elaborated _infoset_,
as opposed to elaborated _infoitem_, has any utility if, for example,
the possibility of embedding a fragment in language B inside a
document in language B, where A and B have different requirements for
quoting vs. elaboration.

The interaction between the two questions arises as follows: are there
actually any generic XML _applications_, that is, applications which
process XML as XML, independently of its vocabulary-specific
semantics, for which a non-application-specific notion of
whole-document elaboration makes sense?  Possible candidates include
validation and style/query processing.  The hard case for this that
emerged in the f2f discussion [1] is that of a vocabulary-specific
conditional construct with an XInclude inside a guarded branch, where
the included material is a very large multimedia document access to
which involves real cost to the user.

This message is an invitation to open up discussion, since answers
evidently are not being offered. . .

ht

[1] http://www.w3.org/2001/tag/2007/03/06-minutes#item10
[2] http://www.w3.org/2001/tag/doc/elabInfoset/
[3] http://www.w3.org/DesignIssues/XML
- -- 
 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)

iD8DBQFHTIlrkjnJixAXWBoRAhJtAJ9cqr2FtTOP0kClHglphAZilka9HACcCLRM
84HZ9zwfdEuwwS0VlhipGCE=
=ERR6
-----END PGP SIGNATURE-----

Received on Tuesday, 27 November 2007 21:17:45 UTC