W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > April 2010

Re: XProc Minutes: 8 Apr 2010

From: murray <murray@muzmo.com>
Date: Wed, 14 Apr 2010 07:19:56 -0800
To: ht@inf.ed.ac.uk (Henry S. Thompson),murray@muzmo.com
Cc: Norman Walsh <ndw@nwalsh.com>,public-xml-processing-model-wg@w3.org
Message-id: <4bc5dd1c.2af.548.25075@muzmo.com>
Henry wrote:
> I guess we ended up aiming _much_ lower.  Or, perhaps,
> tried to answer a different question.
> 
> Question 1 (yours): My XML application requires the
> following processing: x, y, z, p, q, r, . . .  In what
> order should I apply them?
> 

This is the question that I remember us saying we would
address.
I have missed a lot of discussion since then, but I am
surprised
to learn that this question has fallen to the wayside.

> Question 2 (the draft's): My XML application starts from
> an infoset/data-model -- what should I say about how it
> gets built?
> 
> > [C]onceptually, are my expectations too great?
> 
> No, just different.
> 

[...]
> 
> The best I can do for a more principled justification is
> to appeal to two principles to end up at the minimal
> position we've arrived at:
> 
>  1) The operations involved have to be completely
>     application-vocabulary neutral, they address only the
> question of  "what is the information content of this
document";

Puzzled. How do you reconcile this principle with the use of
XInclude?
XInclude has its own vocabulary, as does GRDDL.

According to GRDDL, its processes an infoset which has not 
been augmented by entity or XInclude expansion.
That would seem to be an example of minimal XML processing.
It would seem that this model should be reflected in your
document.

> 
>  2) The operations involved must be fully self-contained
> and automatic, that is, not requiring parameterisation or
> user interaction.

Puzzled again. I guess I missed the discussion that led to
Rule 2.
If the purpose of the model is to describe a process, then I
don't
see what the prtoblem is with describing a process that
happens
to be 'blocked' by user interaction. 
> 
> (1) rules out GRDDL interpretation or xml-stylesheet
> application; (2) rules out decryption and signature
> verification, and arguably validation as well.

Rule 1 does not rule out having to describe the processing
model
that yields the infoset upon which GRDDL-aware processors
operate.
GRDDL does not expect to affect the infoset of the document,
so there is no need to describe or discuss its output.

So far, nobody has spoken directly to the point that I am
raising about
minimal vs expansive XML processing. I understand and
appreciate 
the opinion that minimal XML processing should include the
effects of DTD
contributions, entity expansion and XInclude expansion. That
is the world
in which I think we should be living. Alas and alack, some
XML processors, 
such as GRDDL, do not agree. I think that we should present
at least those
two views of infoset construction. Could one of you please
explain why not?

Thanks,

Murray
Received on Wednesday, 14 April 2010 15:12:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 14 April 2010 15:12:13 GMT