W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2003

AW: Announcement: The "info" URI Scheme

From: Leo Sauermann <leo@gnowsis.com>
Date: Thu, 2 Oct 2003 14:35:26 +0200
To: <www-rdf-interest@w3.org>
Cc: <cowan@mercury.ccil.org>
Message-ID: <000201c388e1$a82f7f10$0b01a8c0@Bundeslade>

> This characterization of the problem is far
> too mild.  Consider the resource whose URI is 
> http://www.tagyerit.com/images/adopted/shakespeare.jpg . You 
> can't do a direct GET on this any more, but 
> it's still available (for now) in the Google cache at 
http://images.google.com/images?q=tbn:IbW4adfRGVkC:www.tagyerit.com/imag
es/adopted/shakespeare.jpg

> If you retrieve this resource, you will get a sequence of bits labeled
image/jpeg.  > 
> But what is the resource?  Is it Shakespeare?  Is it a
> (generic) picture of Shakespeare?  Is it this *particular* image of
Shakespeare?  Or >  
> is it something else altogether?  Nobody can say, probably not even
the person who > 
>> posted the picture to the Web.

Probably that is the mistake here. The person who posts it defines it by
the first triples he associates with the uri.
As we know from Philosophy, terms are correlated to images in our mind
(or something like that), so the nature of the resource is defined by
what it is associated with. Like with a child of yours, you HAVE to
define its name.

If you program using Semantic Web technology, you do
<http://www.tagyerit.com/images/adopted/shakespeare.jpg> <rdf:type>
<foaf:Peson>.

or 
<http://www.tagyerit.com/images/adopted/shakespeare.jpg> <rdf:type>
<foaf:Image>

You are right that you can't guess what the url IS just from the url.
But in Semantic Web the usage of <rdf:type> is here to define what a url
IS.

If you read the Semantic Web primer on RDF you notice, that the uri may
identify BOTH because a uri can be typed with more than one type. It may
not be intelligent but It works very good when you program with it.

I have written hundreds of lines of code where I ask "SELECT ?t WHERE
(<http://blabla/blabla> <rdf:type> ?t)" and get good results.
If a uri supports my desired type, I keep on querying it.
In Jena you write:

Resource r;
r = ....;
if (r.hasProperty(RDF.type, FOAF.Person)) {
	// is a person, do something about it
}


So you don't get "a sequence of bits labeled image/jpeg", that I have
never seen when programming the Semantic Web. You get a RDFS:Class.

Give an example of code where this problem occurs in your applications.

Greetings
Leo Sauermann
Received on Thursday, 2 October 2003 08:36:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:02 GMT