RE: Anonymous resource names -versus- variables

Apologies for the previous truncated message (I just found a new (to me) key
combo in my email client). 

RDF is a language for constructing statements about resources. RDF is often
described in the language of XML. I don't think the problem here is so much
with RDF, or indeed with XML. I think the problem is with the notion of a
resource on the internet. I'm not sure anyone here knows what an internet
resource actually *is* (I certainly don't; any more than I know what photons
actually are), or especially, what delineates a resource from, its name, its
location, its URI, its identity, its dimensionality: or if these things are
indeed separable. My best guess is that a resource is a convenient term for
any and all of these things, and that resource is thus an overloaded term at
best. It's all very interesting stuff, and worthy of argument,  but is
surely getting in the way of working with RDF, since it is difficult to know
what one is making statements about.  

This is compounded by the fact that the internet doesn't have a
corresponding 'real world' for 'maps' to represent into. So, the
map/meta-map analogy fails. But then again the idea a hammer is solid and
not Really Mostly Empty fails too. It *would* be nice if a resource was a
thing, but that doesn't seem very sure. I'm comfortable with a web page (a
page) pertaining to being a 'thing' when I print it out on a piece of paper,
but I'm less sure when I look at it in my browser. By the same token it
would be nice if an RDF resource was really a thing that made statements
about a thing (and not have to frequently print it out to reassure myself).

I humbly suggest that we entertain the group hallucination of resources
being Actual Real Things (ARTs) which are a bit Ghosty And Vague (GAV), and
that these resources can be called referred to and identified by using
Uniform Resource Imaginings (URIs). That way we could pretend indeed that
RDF statements are statements about ARTs (that have URIs), and that these
statements have the interesting aspect that they themselves are ARTs (with
URIs too). Indeed, I put it that this pretence is vital to the success and
widespread adoption of RDF on the web. To borrow a turn of phrase from a
great man whose stock in trade was GAV things: anyone who isn't shocked by
RDF doesn't need to understand it.

-Bill

:-----Original Message-----
:From: Graham Klyne [mailto:GK@dial.pipex.com]
:Sent: 08 May 2000 11:29
:To: jos.deroo.jd@belgium.agfa.com
:Cc: www-rdf-interest@w3.org
:Subject: Re: Anonymous resource names -versus- variables
:
:
:At 02:17 PM 5/5/00 +0200, jos.deroo.jd@belgium.agfa.com wrote:
:>Graham Klyne wrote:
:> > This sparks a thought for me:  RDF recognizes two "layers" 
:of statements:
:> >    statements
:> >    statements about statements, via reification
:> >
:> > Your comments suggest another facet of RDF usage:
:> >    statements about representations of statements
:> >
:> > I sense that some discussion of RDF gets mired because of 
:failure to
:> > distinguish between statements, and some representation 
:thereof.  I regard
:> > this as a vital distinction when considering trust 
:representations (an
:> > immediate concern for CC/PP).
:> >
:> > I regard RDF in terms of some homogeneous collection of 
:statements (and am
:> > thinking about building an experimental system that reflects this
:> > view).  Specific statements are added to the collection 
:via documents that
:> > contain representations --in XML or whatever-- but any 
:meaning associated
:> > with a statement must stand when a particular representation in a
:> > particular document is stripped away.
:>
:>I can feel what you mean and please try to demonstate it!
:>A representation 'stands for something' and
:>a statement is 'something that stands' ???
:
:I wouldn't phrase it quite that way (too many 'stands's ;-)
:
:I'll try and construct an example of what I was saying:
:
:Consider maps:  the real world is continuous, and exists all 
:together.  But 
:a map describes some defined part of the world (both 
:geographically and 
:kinds of features).  One map may show the position of a place 
:with respect 
:to geographical features, and another may show its connections 
:with respect 
:to transport (e.g. street map vs Underground map of London).  
:I think maps 
:are like documents containing (representations of) RDF 
:statements, and the 
:real world is what contains the universe of all the things 
:described by the 
:maps, and more.
:
:Further, imagine that the London Transport authority maintain 
:a master map 
:showing where to find all the instances of Underground Maps 
:displayed in 
:the subway system, and a summary of the features contained on these 
:maps.  This might be thought of as a "statement about 
:representations of 
:statements".
:
:A statement that a map was last updated on a particular date 
:is different 
:from saying something about the place where it is posted.
:
:Much of what we know about the real world may be conveyed by 
:maps, but we 
:need to be able to distinguish between a map and the world it 
:describes.  When we destroy a map, the places it describes continue to 
:exist in the same relationships to each other.  Similarly, I 
:think that the 
:logic (whatever that may be) of some collection of RDF 
:statements should 
:stand in the model and associated schema separately from any 
:document that 
:may contain (some representation of) those statements.
:
:#g
:
:------------
:Graham Klyne
:(GK@ACM.ORG)
:

Received on Monday, 8 May 2000 12:17:50 UTC