W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2001

Re: Datatypes in RDF (was: RDFCore Update)

From: Eric van der Vlist <vdv@dyomedea.com>
Date: Sat, 20 Oct 2001 08:34:46 +0200
Message-ID: <3BD11B06.2060504@dyomedea.com>
To: Patrick.Stickler@nokia.com
Cc: www-rdf-interest@w3.org
Patrick.Stickler@nokia.com wrote:

>>>      <rdf:type rdf:resource="x:dataType"/>
>>Although many vocabularies are doing so I am usually very 
>>reluctant to 
>>use qualified names in attributes or element values.
> The use of qnames as attribute values or in content data aside,
> the above resource is a URV namespace, hence a URI. Yes, I 
> know it looks like a qname. Perhaps it should have been written
> as "x:dataType:". If you haven't already, please read my X-Values
> proposal (I don't want to repeat it here). 
> http://lists.w3.org/Archives/Public/www-rdf-comments/2001OctDec/0088.html

My reluctance about using namespace prefixes is not specific to RDF, 
even though I think that RDF (with RELAX NG mentioned in my previous 
post) is one of the few examples of specs which are sane in this respect.

I have nothing about using prefixes and writing:

<a uri-shortcut="x='http://example.org'" b="x:anything"/>

is something I wouldn't consider as unsane.

It's doing the same with namespace prefixes, especially when this is not 
needed for the markup itself such as in:

<a xmlns:x='http://example.org' b="x:anything"/>

which I consider as unsane since it's creating a dependency from 

the content of a document to its markup.

The namespaces rec acknowledge that a prefix has no value and is only a 
shortcut for a URI. The languages and API acknowledge this as well and 
XPath but also SAX2 and DOM Level2 provide an easy access to the URI but 
*not* to the prefix (and they are clean and right doing so).

> This brings up the question, which I asked in another thread
> yesterday, about whether one can treat a URI scheme prefix
> or URN/URV namespace prefix as a URI.
> E.g. are "http:" or "urn:issn:" URI's? Can I use them to 
> make statements about those schemes? If not, why not?
> (all RDF parsers I've used seem quite happy to treat them as URIs)
> And personally, I think the ability to use qnames in attributes
> is highly desireable and will have a great impact on future user
> acceptance of RDF XML serializations (which are *very* cumbersome
> to write manually -- or require the use of ENTITY tricks to achieve
> the same level of compression and convenience as qnames but 
> introduce yet another representation for a URI). XML Schema got it
> right IMO by allowing qnames as values of URI typed attributes.

I don't think so and I hope RDF will stay "sane" in this respect.

Two specifications (Canonical XML and XPointer) have already hit the 
wall because of this issue and I think it would be high time to 
acknowledge that it's a bad practice :=) ...

My 0.02 Euros


> Cheers,
> Patrick
> --
> Patrick Stickler                      Phone:  +358 3 356 0209
> Senior Research Scientist             Mobile: +358 50 483 9453
> Nokia Research Center                 Fax:    +358 7180 35409
> Visiokatu 1, 33720 Tampere, Finland   Email:  patrick.stickler@nokia.com

Rendez-vous  Paris pour le Forum XML.
Eric van der Vlist       http://xmlfr.org            http://dyomedea.com
http://xsltunit.org      http://4xt.org           http://examplotron.org
Received on Saturday, 20 October 2001 02:34:28 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:44:32 UTC