W3C home > Mailing lists > Public > www-tag@w3.org > May 2007

Re: article on URIs, is this material that can be used by the SWEO IG?

From: Dan Brickley <danbri@danbri.org>
Date: Thu, 31 May 2007 12:50:20 +0100
Message-ID: <465EB67C.5000905@danbri.org>
To: Pat Hayes <phayes@ihmc.us>
CC: John Cowan <cowan@ccil.org>, Leo Sauermann <leo.sauermann@dfki.de>, www-tag@w3.org

Pat Hayes wrote:
> 
>> Pat Hayes scripsit:
>>
>>>  > For example,
>>  > >http://en.wikipedia.org/wiki/Shakespeare is a subject indicator for
>>  > >Shakespeare.
>>>
>>>  But it also mentions Stratford-on-Avon, Mary Arden, and many other
>>>  things. Why is it not just as much 'about' them?
>>
>> Because I don't so employ it.
> 
> OK, fair enough. But then it follows that there is nothing *intrinsic* 
> to that resource that makes it be a subject indicator for Shakespeare. 
> It is so simply because you say it is. But when I read that resource, 
> how do I gain access to *your* intention that it shall be a subject 
> indicator? What readable resource is it that tells me that 
> http://en.wikipedia.org/wiki/Shakespeare is a subject indicator for 
> Shakespeare ?
> 
>>  In principle it could be.  In my judgment,
>> there are other clearly better subject indicators for those resources.
>>
>> Two resources identified by the same subject identifier (or
>> in fact with the same subject-indicator-reference URI)
>> are the same.
> 
> How can a single URI identify two different resources?

I really find the Topic Map approach to this stuff endlessly confusing, 
but perhaps I've just internalised enough of their terminology.

The story to me is simply enough told, just by making this "subject 
indicator" and related concepts into explicit relationships.

  There is a document. It has as its primary topic a Person whose name is
  "Shakespeare". Its URI is http://en.wikipedia.org/wiki/Shakespeare.
  That same document also has other topics, such as the Place
  Stratford-on-Avon, the Person called "Mary Arden", ...

To re-tell this in a machine format, we just need to define the relevant 
relationship types. I put foaf:primaryTopic and foaf:topic into the FOAF 
vocab for just this purpose (and foaf:isPrimaryTopicOf, a redundant 
inverse which can make using Wikipedia URIs syntactically easier in 
RDF/XML). So we have an owl:FunctionalProperty, 
http://xmlns.com/foaf/spec/#term_primaryTopic
"primary topic - The primary topic of some page or document."

This of course doesn't guarantee that all documents it is used with are 
prudently used this way. But the "primary" part does help I think, and 
making that explicit in the OWL can't hurt either. One guideline I've 
toyed with for figuring out which thing is the "primary topic" of a 
document is a simple one: in general, believe the document, if it has 
embedded or extractable RDF/OWL claims. But there are wormcans there too 
if you choose to go looking for them...

>>>  And what does THAT mean? This notion of a thing's "identity" seems to
>>>  be very freely used in W3C circles, but I have no idea what it is
>>>  supposed to mean. What kind of things have an "identity"? Does a
>>>  grain of sand on a beach have an identity? It is certainly identical
>>>  to itself: is that enough?
>>
>> Yes, in principle, but we don't worry about the identity of things
>> which are not the subject of any predications.
> 
> But it is trivial to predicate about grains of sand on a beach. The sand 
> on Pensacola beach, every grain of it, is made of quartz, and so the 
> beach is pure white. There: in the previous sentence I asserted a 
> predication of every grain of sand on a beach. So, do all those grains 
> now have an identity?

Didn't we have this discussion on the RDF lists 6-7 years ago? Back when 
the murkily worded RDF'99 spec and also the URI spec had too much mumbo 
jumbo that tangled up identity, resources, URIs, etc. I thought we'd 
moved on. Some threads are destined to last forever.

>> There are lots of facts we can predicate about Shakespeare; his life
>> is rather well-documented for a person of his place and time.
> 
> True. So what? Does the presence of enough facts make something have an 
> identity? If so, how many facts is enough? What is the threshhold of 
> factual knowledge that creates this mysterious identity-ness?

Quite.

>> Some people assert that some of these facts are wrongly predicated
>> about Shakespeare, like "wrote Hamlet".  Nobody, as far as I know,
>> says that Hamlet was written by some totally unknown person, but
>> if it turned out to be true, that person would be identified as
>> separate from Shakespeare.
> 
> That person would not be Shakespeare. Is that all you mean by your 
> phrasing? Or does your use of "identified as separate" imply something 
> other than simple inequality? (I find this example puzzling, in fact, as 
> presumably if a person is indeed unknown, then surely he or she is not 
> identified at all (??) After all, we can be fairly sure that Shakespeare 
> is not, say, my toaster oven: but that fact about him is not usually 
> thought of as "identifying" him as "separate from" my toaster oven. How 
> does this case differ from your example?)

Random coincidence --- tidying up yesterday I stumbled on an old copy of 
"Shakespeare's plays weren't written by him, but by someone else of the 
same name", http://www.cogsci.indiana.edu/pub/CRCC.publications



> 
>>
>>>  I would feel the same if someone
>>  > were to use the name, say, "Rauschenberg", but about all I know about
>>>  Robert Rauschenberg is that he is, or was, a notable American
>>>  contemporary artist. Is this enough to give him an "identity"? How
>>>  much information about a thing is enough to establish an identity for
>>>  that thing?
>>
>> If there is something you can assert about Rauschenberg, like "is a
>> notable contemporary artist" or even "is named 'Robert Rauschenberg'",
>> then at least you know that he is not Willem de Kooning or Buckminster
>> Fuller or Paganini.
> 
> Why do I know that? For all I know, Willem de Kooning might be a nom de 
> plume for Rauschenberg. But in any case, you havn't answered my 
> question. Is *any* fact about a thing enough to give that thing an 
> identity? It would seem to follow that everything has an identity, since 
> we know some facts about every thing (such as that it is equal to itself).

A strangely comforting thought...

cheers,

Dan
Received on Thursday, 31 May 2007 11:51:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:45 GMT