W3C home > Mailing lists > Public > www-tag@w3.org > April 2008

Re: Announcement: "Cool URIs for the Semantic Web" - W3C SWEO IG Note

From: Golda Velez <gv@btucson.com>
Date: Fri, 11 Apr 2008 11:01:50 -0700
To: Hugh Glaser <hg@ecs.soton.ac.uk>
Cc: Pat Hayes <phayes@ihmc.us>, "www-tag@w3.org" <www-tag@w3.org>, SWIG <semantic-web@w3.org>
Message-Id: <200804111101.50740.gv@btucson.com>

On Friday 11 April 2008 4:30, Hugh Glaser wrote:
> 
> Addressing the first part of Pat's message:
> Thanks Pat, nicely put, and very important.
> It seems to be a common idea that a URI can and even should actually
> identify a non-document resource. In practice this is a fallacy, which leads
> to other broken ideas.

Well - I know I sent this in before, but why not have a vocabulary that 
explicitly uses URIs to identify non-document resources and explicitly says 
what kind of resource they identify?

I made a simple one at http://abra.info/2008/abra-discuss# , it has a class 
called AddressableThing with properties identifiedBy and identifiedAs. 

Now, I can use URIs to identify non-document resources, as the identifedBy 
property of AddressableThing - and I can scoop regular URIs into my system 
using some heuristics and accuracy estimation.  Is this a bad or useless 
thing to do?

<rdfs:Class rdf:ID="AddressableThing"> 
    <!-- would it be useful to be a subclass of daml:Thing ? -->
    <rdfs:comment>Any thing that can be addressed and used as the subject of 
an RDF statement.
	  Should have identifiedBy, identifiedAs and parsingRules properties (the 
latter two may be left as defaults).</rdfs:comment>
  <rdfs:isDefinedBy rdf:resource="http://abra.info/2008/abra-discuss#"/>
</rdfs:Class>

<rdf:Property rdf:ID="identifiedBy">
   <rdfs:label>identifiedBy</rdfs:label>
   <rdfs:comment>When conbined with the defined or default identifiedAs and 
parsingRules, a unique identifier for an AddressableThing.  Usually a URI or 
URN.</rdfs:comment>
   <rdfs:isDefinedBy rdf:resource="http://abra.info/2008/abra-discuss#"/>
</rdf:Property>

<rdf:Property rdf:ID="identifiedAs">
    <rdfs:label>identifiedAs</rdfs:label>
    <rdfs:comment>Is this an RDF statement, electronic document, idea, person, 
place, physical object or other?  Similar to type, but means 'what kind of 
thing is identified by this resource'.  Defaults to document for class 
AddressableThing, and to RDF statement for the class 
AddressableStatement.</rdfs:comment> <!-- TODO here: needs range -->
</rdf:Property> 



> Such URIs do not properly identify anything any more than words do in
> language. If we expected words to be such accurate identifiers, we would be
> unable to begin non-trivial communication; if we expect URIs to be such
> accurate identifiers, we will be unable to build a non-trivial Semantic Web.
> 
> Best
> Hugh
> 
> On 09/04/2008 06:12, "Pat Hayes" <phayes@ihmc.us> wrote:
> 
> > Nice document. A few quibbles:
> >
> > -----------
> >
> > "Given such a URI, how can we find out what it identifies? We need some 
way to
> > answer this question, because otherwise it will be hard to achieve
> > interoperability between independent information systems. "
> >
> > I know you probably don't want to get involved with philosophical issues, 
but
> > this sentence is so wrong as to be misleading. We CANNOT POSSIBLY find out
> > what any name identifies, other than by being explicitly handed the thing 
or
> > being pointed to it (as in the pre-semantic Web). The best we can POSSIBLY 
do
> > is to have a detailed enough description of the thing for our purposes; 
but no
> > description can completely identify a single thing. So your second 
sentence
> > above is particularly misleading: it suggests that the Web will not work
> > unless we all do something impossible before breakfast.
> >
> > Moreover, getting involved with this highly debatable issue isn't needed, 
to
> > motivate the subsequent content of the note. The point is not that we need 
to
> > be able to discover what exactly it is that non-document URIs denote. The
> > central point is only that, pretty much by definition, they denote 
something
> > that isn't a document, which is why we need to distinguish the URI for the
> > thing from a URI for any document/"information resource" which describes 
the
> > thing. You can avoid the philosophical/semantic tar-pit of the nature of
> > reference and how it can be determined, by sticking to this basic point in
> > your introduction.
> >
> >
> > ------------
> >
> > Best wishes
> >
> > Pat Hayes
> 
> 

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Golda Velez	gv@btucson.com		520-440-1420
	 	Tucson Superblog	http://btucson.com
		Search software		http://webglimpse.net
	Say Anything About Anything	http://abra.info
Received on Friday, 11 April 2008 17:45:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:55 GMT