W3C home > Mailing lists > Public > www-rdf-comments@w3.org > July to September 1999

Re:Description vs. Class

From: astrid storr <Storr@t-online.de>
Date: Tue, 27 Jul 1999 15:04:32 +0200
Message-Id: <m1196uG-0007L6C@fwd14.btx.dtag.de>
To: rdf-dev@mailbase.ac.uk
Cc: www-rdf-comments@w3.org
Hi everybody,

On Tue, 20 Jul 1999, Olivier Marce wrote:
>Exact: using rdf:ID (and *not* rdfs:ID as erroneously stated previously) 
>gives the correct (?) two triples. 
>
>Anyway, since a <rdf:Description rdf:ID="foo"> seems to be different
>than <rdf:Description ID="foo">, it is not clear (for me at least) if
>the recognized identifier attributes must belongs to the 'rdf'
>namespace (as implemented in SiRPAC) or to the local namespace of the
>element. The last would mean that we shoud use 'ID' attribute name
>instead of 'rdf:ID'.

I think he hit the problem. The following attributes used in the 
RDF M&S:
	about, aboutEach, aboutEachPrefix, bagID, ID, resource
	and parseType
These attributes are not described in the RDF-NS (http://www.w3.org/
1999/02/22-rdf-syntax-ns). 

If those attributes belong to the RDF-NS, I would say
<rdf:Description rdf:ID="foo"> and <rdf:Description ID="foo"> are equal
because in the second the ID-Attribute inherents the RDF-NS. But some-
thing like <Description ID="foo"> with the RDF-NS as default would be
different, refering to 'Namespaces in XML 5.2 ... Note that default 
namespaces do not apply directly to attributes.' 
And then there are also some errors in the RDF-Schema-Spec. 'Compact XML
Serialization', where we can find: <rdfs:Class about="http:...#Statement">.

If those attributes belong to the local namespace of the element then
<rdf:Description ID="foo"> and <Description ID="foo"> with the RDF-NS as 
default would be equal and <rdf:Description rdf:ID="foo"> would mean 
something different!?! 
Then we also could have some errors in the RDF-NS, where you can find some-
thing like: <s:Class rdf:ID="Bag" ...>.


Well, I would prefer that those attributes belong to the RDF-NS and I think 
the most examples are written in that way. If not those names should not
be used as ID-Names, I think otherwise this could occour new problems.


Lets come back to the statement:"Note that default namespaces do not 
apply directly to attributes." What does it mean exactly? 
I would say it just mean the following: If we have
	... xmlns="NS-A"
	    xmlns:B="NS-B" ...
	<B:Name Attribut="...
that in this case the attribute "Attribut" does not belong to the default
NS. And the NS is inherent by the Element-NS, but if we have:
	... xmlns="NS-A"
	    xmlns:B="NS-B" ...
	<Name Attribut="...
the attribute "Attribut" belongs to the default NS inherent by the
Element-NS. (So, the default NS apply indirect to this attribute!)
Am I wrong??????


If I am not wrong and if attributes like ID belong to the RDF-NS,
I would say that <rdf:Description rdf:ID="foo">, 
<rdf:Description ID="foo"> and <Description ID="foo"> 
(with RDF-NS as default) are all three the same!


Wether I am wrong or right, I think the  'Compact XML Serialization' 
in the RDF-Schema-Specification is not consistent, because you can find:
<rdfs:Class rdf:ID="Literal" ..... and
<rdfs:Class about="http...#Statement> ...
I do not see a reason why "ID" and "about" should be treated in different
ways.

Regards,
Karsten Tolle (tolle@csi.forth.gr)
-----------------------------
Received on Tuesday, 27 July 1999 09:05:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:27 GMT