W3C home > Mailing lists > Public > www-rdf-interest@w3.org > March 2000

Re: Subclass of Thing/Resource

From: Tim Berners-Lee <timbl@w3.org>
Date: Wed, 1 Mar 2000 16:25:50 -0500
Message-ID: <010f01bf83c4$c4fce540$a60a1712@col.w3.org>
To: "Guha" <guha@epinions-inc.com>
Cc: "www-rdf-interest" <www-rdf-interest@w3.org>
(I have added this - as it was getting a bit long - to my nurdling son the problems of identity in RDF in http://www.w3.org/DesignIssues/Identity)

Guha,

I take the word "resource" to be synonymous with "that identified by a URI". 
This is generally something which is referred to in a protocol. The semantics are defiend by the protocol spec.
E.g. an Mailbox is a paarmeter to an SMTP operation, an object which can appear in a To: field, etc.

HTTP defines the reationship between requests (which include URIs) and responses (which include strings of BITS and MIME types and things).   An HTTP URI represents an abstract doucment of which you can get (sometimes) representations using HTTP. An md5: URI represents  strings of bits.

Not all things have URIs.  For example, URIs. For exaple, I don't have a URI. 

Maybe RDF allows you to generate a URI for any concept. This is not clear. It is not clear for the reason which follows, and also for the reason that literals don't have URIs because the data: URI gives a MIME type and not and XML type. But we can get around the latter.

Ralph and I yesterday discussed what is identified by http://.../foo.rdf#bar when ...foo.rdf contains:

<rdf:description rdf:id="bar">
   <rdf:type resource="...#person">
   <y:common-name>Ora Lassila</y:common-name>
   <y:mailbox>ora.lassila@research.nokia.com</y:mailbox>
</rdf:description>

It seems taht the implication is that if this is takes as a document of type application/rdf, then the gragment identifier identifies the thing (person in this case, Ora) described RDF node.

However, if its considered to be of type text/xml then the fragment identifier is defined bythe XML spec, and so references soemthing with type XML:ID. and value "bar".   It happens that the rdf:id is not defined to be an xml:id but is defined to act like one, whatever that means, by the RDF spec. So it isn't cl;ear whether the reference to this would be to the XML subtree (consisting of the rdf:description element and its contents) or would be undefined or possibly a refernce to some other element which happened to have id="bar".

I feel that this needs to be sorted out.  It is not enough just to let XML as a language refer to the subtree (eg in a link), and let RDF always be defined to refer to the object described by the RDF ndoe described by the subtree. This is because RDF needs to be able to make statements about XML documents and subdocuments. So for example, I might want to state that I wrote the above. It would be very tempting to write that I am the author of foo.rdf#bar. But I am not the author of Ora Lassila.

So, at the moment RDF uses and parseType to resolve this for inline data.

The thing could be resolved with an intyerpretive property which expressive relationship between an XML subtree and an RDF object which it describes.  But it needs a shortcut.

I would propose that "resource=" which is used to point to a resource be also used for a resource fragment id, and that a new syntax be introduced to refer to the actual RDF node.  maybe "object=" which happens here to correspond to the (subject, predicate, object) sense -- as well as a "thing" sense.  (The former is what is the reason for chosing it - the attribute should epxtrezs therelationship, not the class of the thing refered to in general!).

Tim
    -----Original Message-----
    From: Guha <guha@epinions-inc.com>
    To: Dan Brickley <Daniel.Brickley@bristol.ac.uk>
    Cc: Tim Berners-Lee <timbl@w3.org>; www-rdf-interest <www-rdf-interest@w3.org>
    Date: Wednesday, March 01, 2000 12:35 AM
    Subject: Re: Subclass of Thing/Resource
    
    
    I remember a heated discussion about this very issue. 
    Tim, if you do think there are things that cannot be resources 
    (in the URI sense of the word), I'd really like to know what 
    those are! 

    guha 

    Dan Brickley wrote: 

        On Tue, 29 Feb 2000, Tim Berners-Lee wrote: 
        > In RDF, every class is a subclass of rdfs:Resource so surely that is just as 
        > informationless too. 
        > (I do wish RDF had used "thing" instead of "resource" which has a meaning in 
        > URI already). 
        > 
        > TimBL 
        So... are there identifiable 'things' that aren't 'resources'? (do you 
        mean the Literals?). The RDFS design on this pretty much followed from the RDF 
        Model claim that we inhabit the Web's world of URI-nameable resources... 

        Could you say something about what you'd count as a non-resource 'thing'? 
        (or is it a concern that we've coupled the RDF terms to concepts 
        defined, and argued about, elsewhere...?) 

        --dan
Received on Wednesday, 1 March 2000 16:27:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:42 GMT