W3C home > Mailing lists > Public > public-xml-core-wg@w3.org > September 2004

FW: Transition Request: PR Request for XInclude

From: Paul Grosso <pgrosso@arbortext.com>
Date: Tue, 14 Sep 2004 14:08:53 -0400
Message-ID: <F13E1BF26B19BA40AF3C0DE7D4DA0C03783625@ati-mail01.arbortext.local>
To: "XML Core WG" <public-xml-core-wg@w3.org>

Our XInclude PR request generated some discussion.

I am forwarding excerpts to the WG list for answers (see
http://lists.w3.org/Archives/Member/chairs/2004JulSep/0090
for the full thread if you wish).

I am being asked for answers to all these issues.

paul

----- START Original Message #1-----
From: Karl Dubost [mailto:karl@w3.org] 
Sent: Tuesday, 2004 September 14 00:57

Le 14 sept. 2004, à 01:24, Paul Grosso a écrit :
> The test suite for XInclude is available at:
> http://www.w3.org/XML/Test/XInclude/
>
> The implementation results are available at:
> http://www.w3.org/XML/2004/xinclude-implementation/report.html
>
> The results show two or three implementations of all required features
> and at least one implementation of the above two mentioned optional
> features; therefore, we have fulfilled our CR exit criteria.

 From the table, I can see that the features teste by Nist-include-xx, 
where xx is
	19, 24, 25, 48, 49, 51, 52, 56
	and eduni-3

are at risks. Are those features optional?
----- END Original Message #1-----


-----START Original Message #2-----
From: Martin Duerst [mailto:duerst@w3.org] 
Sent: Tuesday, 2004 September 14 01:59

It would be very good if that table contained pointers to the
implementations and to the tests.

Also, when searching the test description
(http://dev.w3.org/cvsweb/~checkout~/2001/XInclude-Test-Suite/testdescr.xml? 
rev=1.14&content-type=text/xml) for
'IRI', the only thing I found was:

<!--
  (RMT) Removed because we can't reliably put non-ascii characters in
            file names:

                 <testcase id="Nist-include-57" href="nist-include-57.xml"
                           type="error">
                         <contributor>Sandra I. Martinez</contributor>
                         <section 
resource="http://www.w3.org/XML/Group/2003/03/PR-xinclude-20030331/#IRIs">5. 
0</section>
                         <date qualifier="created">April, 2003</date>
                         <description>  Simple IRI test. </description>
                 </testcase>
   -->

I agree that it is not possible to produce a test suite that can be
deployed locally on a variety of systems and would guarantee the
correct characters (in the correct encoding) in file names.
However, this should not mean that such tests should not be produced.
There are at least two ways in which such tests can be produced and run:
- Create different versions of the test suite for different platforms
   (.tar.gz would include the file names in UTF-8 for use on Unix and
    MacOS X, and we would have to test/check zip to see whether it does
    the right thing with file names on Win NT/2000/XP, and otherwise
    would have to look for another tool for these systems (e.g. some
    kind of installer), and would have to very clearly indicate the
    limits of these tests (maybe by also adding some negative tests
    with a high possibility of turning positive in case of setup problems)
- Create a careful setup at www.w3.org. I can easily help with this.


I would be very glad to work together on this,
it shouldn't be a big deal.
----- END Original Message #2-----


-----START Original Message #3-----
From: Karl Dubost [mailto:karl@w3.org] 
Sent: Tuesday, 2004 September 14 02:26
To: Karl Dubost
Cc: w3t-qa@w3.org; w3t-comm@w3.org; chairs@w3.org; webreq@w3.org; Steve Bratt; Paul Grosso; w3c-xml-cg; timbl@w3.org; ht@w3.org; Jonathan Marsh
Subject: Re: Transition Request: PR Request for XInclude

Another possible problem

XInclude relies on IRI
	http://www.w3.org/TR/2004/CR-xinclude-20040413/#IRIs

but IRI is not yet a specification.	
	http://www.w3.org/International/iri-edit/draft-duerst-iri-02.txt

The Specification then tries to define its own definition of IRI, which 
might lead to mistakes if the IRI draft changes for one reason or 
another. It's dangerous and might lead to incompatibilities.

What would you suggest?
----- END Original Message #3-----

Then there was more follow up by Martin Duerst and others
about the state of play of the IRI spec and the fact that
we have some errors in the XInclude spec in the IRI area.

My ISP is having problems today, so I can't easily give
all the pointers, but check the chairs archive for details.
Received on Tuesday, 14 September 2004 18:11:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:31 GMT