W3C home > Mailing lists > Public > public-xml-core-wg@w3.org > December 2007

RE: I-D Action:draft-duerst-iri-bis-02.txt

From: Grosso, Paul <pgrosso@ptc.com>
Date: Fri, 21 Dec 2007 11:11:32 -0500
Message-ID: <CF83BAA719FD2C439D25CBB1C9D1D30209D17722@HQ-MAIL4.ptcnet.ptc.com>
To: "Martin Duerst" <duerst@it.aoyama.ac.jp>, <public-iri@w3.org>
Cc: <public-i18n-core@w3.org>, <public-xml-core-wg@w3.org>, "Addison Phillips" <addison@yahoo-inc.com>

Martin,

Thank you for the detailed response.

Some comments below. 

> -----Original Message-----
> From: public-xml-core-wg-request@w3.org 
> [mailto:public-xml-core-wg-request@w3.org] On Behalf Of Martin Duerst
> Sent: Friday, 2007 December 21 0:13
> To: Grosso, Paul; public-iri@w3.org
> Cc: public-i18n-core@w3.org; public-xml-core-wg@w3.org; 
> Addison Phillips
> Subject: RE: I-D Action:draft-duerst-iri-bis-02.txt 

> Here are the things that others, including the XML Core WG,
> may be able to help:
> - Provide tests, run the tests and create test matrixes.
>   This would be particularly helpful for the LEIRI case;
>   it's much easier to make the case of the LEIRI section
>   if we can show that actually existing XML implementations
>   actually accept LEIRIs that are not IRIs

Our specs are not yet Proposed Edited Recommendations (PERs)
which, as you know, is similar to the Candidate Recommendation
(CR) stage in which one requests implementation experience.

On the other hand, we have made sure the definition of LEIRIs
matches the current wording in the XML spec, so it should be
the case that most implementations now handle LEIRIs.

Given the status of iri-bis described in your message, I am
now considering taking our various specs to PER in January
(though I would need the XML Core WG's agreement and the
cooperation of the various editors, of course) referencing
your iri-bis-02 (or, if available, later) draft from the PERs.

We would be able to tabulate implementation experience, and
we would plan to exit PER and go to Rec when the iri-bis
is a referenceable RFC.

>   For XML system identifiers, I can easily create a test 
>   myself, it doesn't
>   need more than an XML document with a few elements and a 
>   few lines of DTD.
>   However, for the other specs (you mention a total of six), 
>   I'd really
>   appreciate if you could create the tests, or send me some
>   if they already exist (e.g. for XML Schema).

XML Schema is not one of our specs.

Aside from XML 1.0, I wouldn't think the rest of our specs  
need independent testing, though perhaps some test cases 
can be tailored to test specific areas in such specs.

In particular, we are talking about the following:

XML 1.0 6th (or 5th) Edition PER
--------------------------------
We are planning to produce a 5th Edition with the change
to the name characters first, but given your latest status
report, the WG may decide to combine the LEIRI change with
the 5th Edition.  Our concern is that we didn't want the
name char change to be held up waiting for the IRI RFC.

XML 1.1 4th (or 3rd) Edition PER
--------------------------------
This needs no special LEIRI tests beyond those for XML 1.0.
Again, whether the LEIRI change ends up in the 3rd or 4th
Edition depends on whether we decide to combine the LEIRI
change with other changes.

XML Base 2nd Edition 2nd PER
----------------------------
The current status of XML Base is the 20 December 2006 PER.
We plan to re-attempt the PER with the LEIRI change.

Other than ensuring that there are some tests in the XML 1.0
suite that have LEIRIs as the value of the xml:base attribute,
no other special testing is needed for XML Base.

XLink 1.1 (First Edition) - next stage unclear
----------------------------------------------
The current status of XLink 1.1 is the 28 March 2006 CR.
We had what one could consider a successful (enough) CR,
but we have been waiting to be able to address what we
are now calling the LEIRI issue before going to PR.

It isn't clear whether we can just put in the LEIRI text
and go to PR or not.  If not, then I believe we have to
go back to Last Call.  I don't see any need for any tests
specific to XLink beyond those we do for XML 1.0.

XInclude 3rd Edition PER
------------------------
Other than ensuring that there are some tests in the XML 1.0
suite that have LEIRIs as the value of the xinclude's href 
attribute, no other special testing is needed for XInclude.

Infoset
-------
It turns out this references XML Base and XML, but does not
itself need any changes related to LEIRIs, so in fact we do
not plan to reissue this.

paul
Received on Friday, 21 December 2007 16:11:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:37 GMT