W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > January 2009

Re: Validation issue with the xsi namespace

From: Shane McCarron <shane@aptest.com>
Date: Thu, 22 Jan 2009 05:59:08 -0600
Message-ID: <49785F8C.8050505@aptest.com>
To: Christoph LANGE <ch.lange@jacobs-university.de>
CC: Stephane Corlosquet <stephane.corlosquet@deri.org>, public-rdf-in-xhtml-tf@w3.org

It basically has to do with dereferencing the resulting URI when you use 
a QName or CURIE (e.g., xsi:lala should dereference to 
http://www.w3.org/2001/XMLSchema-instance#lala - basically it has to do 
with how the resource at the end of the namespace URI is constructed, 
and how its components are accessed.  If the resource masquerades as a 
folder, then http://www.w3.org/2001/XMLSchema-instance/lala will 
magically return the description for lala.  If it masquerades as a 
document, then http://www.w3.org/2001/XMLSchema-instance#lala will 
return the complete resource and there will be a ID within that resource 
that corresponds to lala.

CURIE vocabularies tend to use the # convention, FWIW.  In this case, 
xsi is NOT a CURIE vocabulary, so I would use whatever namespace URI is 
defined by the xsi specification.

Christoph LANGE wrote:
> Hi Stéphane,
> 
> On Thursday 22 January 2009 12:42:07 Stephane Corlosquet wrote:
>> Line 3, Column 55: value of fixed attribute "xmlns:xsi" not equal to
>> default.
>>   xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance#"
>>
>> while xmlns:xsd does not produce such an error.
>>
>> Removing the # in the end of the xsi prefix fixes the error, but I
>> wonder if it's correct to do so to see prefixes finishing by / or #.
> 
> Sure, prefixes of XML element namespaces usually do not end with / nor #.  The
> xsi namespace is not such a namespace, but the whole context of XML Schema is
> rather related to XML elements than to semantic web vocabularies, and that's
> why I assume they adopted that convention.
> 
> On the other hand I've never understood _why_ these different conventions
> exist.  Hash vs. slash for ontologies has been discussed, but I still have not
> seen a survey and discussion of "hash/slash vs. nothing". -- Does anybody know
> more?
> 
> Cheers,
> 
> Christoph
> 

-- 
Shane P. McCarron                          Phone: +1 763 786-8160 x120
Managing Director                            Fax: +1 763 786-8180
ApTest Minnesota                            Inet: shane@aptest.com
Received on Thursday, 22 January 2009 12:00:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 22 January 2009 12:00:07 GMT