RE: FW: XInclude use case for Device Independence

I gather from this brief interaction that AVTs are an appropriate
approach. Given that XInclude does not (currently) support AVTs, we in
DIWG should consider two possibilities:
1. A future XInclude that will support AVTs
2. Enhancing DISelect to support AVTs

We had noted the issue of the processing model. We have similar issues
when we consider what happens if DISelect is chained (possibly via
XInclude), and also what happens if re-evaluation is necessary following
some event that changes the state upon which DISelect is executing.

A third possibility, suggested by T.V.Raman during this interaction, is
the general extension of XML to support AVTs. No doubt this would be, to
use his own words, "a huge win" but I am sure that there would be many
implications for XML that are far beyond the remit of DIWG to consider.
We would, nevertheless, support any investigation in this area that the
XML Core WG would consider undertaking.

Regarding use cases, let me just note that the coupling of DISelect and
XInclude provides a markup-based model for portal/aggregation web content.
Given the nature of DISelect, this model is ideally suited to mobile and
other non-traditional (i.e. non-desktop) web content solutions.

Finally, please note that DISelect is still in its early stages and while
we are still able to add features and functionality to the technology, it
is the general opinion of the group that such additions should be kept to
a minimum. If there is a possiblity that XInclude or XML itself would adopt
the features of AVTs, then the necessity to amend DISelect would be
avoided. While this would obviously place a burden on XML Core, it would
probably be the wiser course of action in the long term.

Regards,
---Rotan.

-----Original Message-----
From: Norman Walsh [mailto:Norman.Walsh@Sun.COM]
Sent: 21 December 2004 16:27
To: public-xml-core-wg@w3.org
Cc: Rotan Hanrahan
Subject: Re: FW: XInclude use case for Device Independence


/ Rotan Hanrahan <Rotan.Hanrahan@MobileAware.com> was heard to say:
| Thank you Norman. I am copying to DIWG. Your mention of XSLT is in keeping
| with a recent update to this issue in which "Attribute Value Templates" is
| proposed for consideration. I'm sure we shall return to this complex issue
| after the holidays.

Yes, I had been made aware of that and didn't notice that it wasn't
mentioned in the message I replied to.

I think using AVTs fits in with one of the suggestions I made:

| -----Original Message-----
| From: Norman Walsh [mailto:Norman.Walsh@Sun.COM]
| Sent: 21 December 2004 16:01
| To: public-xml-core-wg@w3.org
| Cc: Rotan Hanrahan
[...]
| Alternatively, you could invent an expression language and say that
| your processing model is: load the document with XInclude suppressed,
| evaluate the expressions, and then apply an XInclude processor.

Instead of inventing an expression language, you could say that you
expand AVTs. You still have to fiddle with the processing model, of
course, because you don't want the processor to attempt to expand
XInclude before you evaluate the AVT!

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@Sun.COM / XML Standards Architect / Sun Microsystems, Inc.
NOTICE: This email message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution is prohibited.
If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.

Received on Tuesday, 21 December 2004 17:16:01 UTC