Re: Contexts (spinoff from copy and wrap rdf statements)

I think I'd better make a few remarks about the difference between parts and attributes.

By definition, a part of an entity is something that can be physically separated from the entity.  When we separate a part, that part becomes a "new" entity.

So, the question is: is a statement in a document a "part" which can (will) be removed?  The answer is not immediately obvious to me.  For example, we could forbid editing, i.e., we could require a document to be completely replaced by a new document.  When we build a context consisting of more than one document, will we allow the use of a single statement, excluding the rest of the document?  When a document changes, do we want to talk about which statements were changed?

First, the above questions need to be answered; then we can see if a statement is a "part" or an "attribute".
============ 
Dick McCullough 
knowledge := man do identify od existent done
knowledge haspart list of proposition

  ----- Original Message ----- 
  From: Richard H. McCullough 
  To: seth@robustai.net 
  Cc: Danny Ayers ; David Menendez ; rdfig 
  Sent: Saturday, November 23, 2002 1:45 PM
  Subject: Re: Contexts (spinoff from copy and wrap rdf statements)


  You are absolutely right.
  Pretending a "part" is an "attribute" may cause trouble as we try to use it.

  I'm saying: from my experience, it's easier to work with attributes than with parts.
  Maybe it's time for me to do some more work with parts.
  ============ 
  Dick McCullough 
  knowledge := man do identify od existent done
  knowledge haspart list of proposition

    ----- Original Message ----- 
    From: Seth Russell 
    To: Richard H. McCullough 
    Cc: Danny Ayers ; David Menendez ; rdfig 
    Sent: Saturday, November 23, 2002 12:06 PM
    Subject: Re: Contexts (spinoff from copy and wrap rdf statements)


    Richard H. McCullough wrote:

      I would consider the set of statements in a document (or graph) to be a property/value of the document (or possibly a "part", but I think that's an unnecessarily complicated viewpoint).  Now you can talk about that property/value, define a truth-value property for it, etc.
    I don't think it is a property in the sense of rdfs:Property.     But a statement is certainly a part of a document, and a triple is certainly a part of a graph.  I don't see any reason we couldn't assign truth values to statements that talk about these things... formally:

    language: Semenglish
    {<foo.rdf> docContains "A r B."} entails {<foo.rdf#ThisGraph> graphContains {A r B}}. 
    ThisEmail author (Seth Russell). 
    (Semenglish Primer) seeUrl <http://robustai.net/mentography/semenglish.html>.



      ============ 
      Dick McCullough 
      knowledge := man do identify od existent done
      knowledge haspart list of proposition

        ----- Original Message ----- 
        From: Danny Ayers 
        To: Richard H. McCullough ; David Menendez ; rdfig 
        Sent: Saturday, November 23, 2002 3:53 AM
        Subject: RE: Contexts (spinoff from copy and wrap rdf statements)




         
          If you let a resource refer to itself, you can just say
              resource has
                  graph = "...",
                  document = "..."
          (however you want to say it in RDFS)
          so the graph would have a reference to itself and the document,
          and ditto for the document.  

          Having such a "cross-reference" doesn't cause any problems, does it? 

          Probably not. 

          Aren't the graph and document "isomorphic", i.e., logically equivalent, or
          are you talking about a different kind of document here?

          Hmm - that's the crunch I suppose. A HTML document can be a resource and have a URL that can be used as its URI. But do we consider an RDF document in the same circumstances a closed box, or a bunch of 'free' statements..? Similarly, if the HTML doc (let's make that XHTML+XLink) made RDF-friendly statements ("myMetaDataHere: me.rdf") how available to the referrer should those statements (and anything else they refer to), be? 

          I guess this is back into the "dark triples" idea.

          If statements are directly asserted by this then they lose their provenence, if they are quoted/reified then that brings up the question of unquoting/unreification mechanisms.
          Hmm... 

          Cheers,
          Danny.