Re: Action Item: http://www.w3.org/2006/04/19-i18nits-minutes.html#action01 (handling of inclusions)

Hi Yves, all,

Yves Savourel wrote:
> Hi Christian, Sebastian, all,
> 
>> <note>
>> The XPath expressions used by ITS selection assume that any inclusions 
>> (such as those based on XInclude) are resolved before selection is 
>> applied. Accordingly, inclusion mechanisms such as XInclude or DITA's 
>> <gi>conref</gi> may need to be followed before ITS selections are 
>> applied. </note>
> 
> Mmm... I thought we had reached the opposite conclusion (?)

I think so too. Maybe the wording is a little bit ambigous? Christian,
Sebastian, do you agree with Yves interpretation?

At
http://www.w3.org/International/its/itstagset/itstagset.html#def-selector
, we have

[[Selection relies on the information which is given in the XML
Information Set [XML Infoset]. ITS applications may implement inclusion
mechanisms such as XInclude or DITA's [Dita 1.0] conref.]]

which I wrote during the f2f, having Yves interpretation in mind.


> 
> I thought we would have something more like: "The ITS selection mechanism applies to the PSVI of the document. It is assumed that
> inclusion mechanisms such as XInclude of DITA's conref are not taken in account when ITS selections are applied."
> 
> In the current wording I see the "may need" but the "assume... are resolved before selection" seems to indicate that a processor
> does not have really a choice, and that 'may need' is really a 'must'.
> 
> It would be a bit un-realistic to ask all ITS processors (e.g. generic XML translation systems) to implement every inclusion
> mechanism known (and the future ones...).
> 
> Another solution, as Sebastian proposed during the face-to-face, would be to not say anything, or say ITS does not care: it's a tool
> choice. But this would open the door for interoperability issues. We do not want that.
> 
> How other application handle inclusion? We can't compare ITS to CSS here because CSS is applied at the very end of the whole
> process, while ITS is applied at the beginning. XSLT would ignore inclusions, wouldn't it?

Both XSLT and XQuery operate on a datamodel which is an extension of the
 XML Infoset. These specs define what is in this data model, but not how
it is constructed, see
http://www.w3.org/TR/xquery/#id-data-model-generation :
[[An XDM instance might also be synthesized directly from a relational
database, or constructed in some other way (see DM3 in Fig. 1.) XQuery
is defined in terms of the data model, but it does not place any
constraints on how XDM instances are constructed.]]

So these specs do not say nothing, but say "you can do everything to
create a data model instance, but it is your responsibility, and we
don't care."

I don't know how to word that, though ...

> 
> Cheers,
> -yves
> 
>