Re: impossible (a URI is a name)

Whoops!  It's not impossible.  We could make the type name itself a component of the URI.  

But that's NOT a good idea.  The type names change with the Ontology, 
and the Ontology changes with the context of "what is the purpose?".
============ 
Dick McCullough 
knowledge := man do identify od existent done
knowledge haspart list of proposition

  ----- Original Message ----- 
  From: Richard H. McCullough 
  To: Paul Prescod 
  Cc: Doug Ransom ; www-rdf-interest@w3.org 
  Sent: Sunday, November 24, 2002 3:36 PM
  Subject: Re: a URI is a name (tel uri scheme and VCARD RDF)


  I agree with you.

  Viewing a URI as a name, it's impossible to use the lexical structure of the name to signal the type of the thing referred  to -- Person, Graph, Document, etc.  There are too many different types.

  The syntax structure does help, e.g., the triple

      subject  property  object

  conveys some information concerning the types of the referents.
  ============ 
  Dick McCullough 
  knowledge := man do identify od existent done
  knowledge haspart list of proposition

    ----- Original Message ----- 
    From: Paul Prescod 
    To: Richard H. McCullough 
    Cc: Doug Ransom ; www-rdf-interest@w3.org 
    Sent: Sunday, November 24, 2002 2:48 PM
    Subject: Re: a URI is a name (tel uri scheme and VCARD RDF)


    Richard H. McCullough wrote:

    > I followed the link at the bottom of this email, and read Sean B. 
    > Palmer's interesting essay on URIs. 

    I'm not clear on your opinion of Doug's proposal. ;) But you did change 
    the subject line so I guess we'll treat this as a new thread.

    > For example, if we want to talk about Linus Torvalds in the context of 
    > URIs & RDFS, we might have three different somethings that we need 
    > three different names for:
    >
    >     Linus Torvalds the person
    >     a document that contains RDFS statements that describe facts about 
    > Linus Torvalds
    >     a graph that contains nodes and links that describe facts about 
    > Linus Torvalds
    >
    > My impression, based on a few RDF-interest emails, is that much 
    > confusion has been generated by trying to use one URI when three URIs 
    > are required. 

    I think that you misunderstand the debate. It is easy to make up three 
    URIs for Linus Torvalds or ten (let's not forget his home page and email 
    address and ...). The question is whether the syntactic form of the URI 
    restricts whether it refers to him, or his home page or the graph or ...

    Seth says:

     >  Linus Torvalda the person
     >    http://foo/#LinusTorvalds
     >  a documdent that contains RDFS statements that describe facts about
     > Linus Torvalds
     >    http://foo/
     >  a graph that contains nodes and links that describe facts about 
    Linus > Torvalds
     >    http://foo/#ThisGraph

    But an equally consistent position is:

    Linus Torvalds the person

         http://foo/LinusTorvalds

       a docudent that contains RDFS statements that describe facts about 
    Linus Torvalds

         http://foo/LinusTorvaldsInRDF

       a graph that contains nodes and links that describe facts about Linus 
    Torvalds

         http://foo/LinusTorvaldsGraph

    Insofar as RDF care NEITHER about the syntax of the URI _nor_ the data 
    referred to by the URI, why should WE care? Why impose a syntactic 
    convention at all? RDF offers me ways of saying that InRDF is the RDF 
    representation and Graph is the graph representation of the 
    LinusTorvalds concept. If Seth wants to use his convention then he can, 
    but he should also use RDF statements to make explicit the relationship.

      Paul Prescod

Received on Sunday, 24 November 2002 19:06:33 UTC