W3C home > Mailing lists > Public > www-tag@w3.org > July 2003

RE: resources and URIs

From: pat hayes <phayes@ihmc.us>
Date: Mon, 21 Jul 2003 17:25:57 -0500
Message-Id: <p06001a05bb4214e2314a@[]>
To: "Williams, Stuart" <skw@hp.com>
Cc: www-tag@w3.org

>  > -----Original Message-----
>>  From: pat hayes [mailto:phayes@ihmc.us]
>>  Sent: 19 July 2003 19:42
>>  To: Williams, Stuart
>>  Cc: www-tag@w3.org
>>  Subject: RE: resources and URIs
>>  >Tim (or Pat),
>>  >>  No, that would be illegal by my way of thinking. 
>>  >> http://chandra.harvard.edu/NGC/ngc1068 is an information resource. 
>>  >> You would expect
>>  >>
>>  >>
>>  >>  <rdf:Description 
>>  rdf:about="http://chandra.harvard.edu/NGC#ngc1068"
>>  >>  rdf:type="http://chandra.harvard.edu/AOtype/Activegalaxy7"
>>  >>  </rdf:Description>
>>  >>
>>  >>  or, in the http://chandra.harvard.edu/NGC information ressource,,
>>  >>
>>  >>  <rdf:Description
>>  >>  rdf:about="#ngc1068" 
>>  >> rdf:type="http://chandra.harvard.edu/AOtype/Activegalaxy7"
>>  >>  </rdf:Description>
>>  >>
>>  >>  where you can see that local identifiers can be used to refer  to
>>  >> abstract things, because that is what the RDF language spec says.
>>  >
>>  >Can you provide a reference to what the RDF language spec actually says
>>  >on this topic, I'd like to read it for myself.
>>  The relevant citation is probably
>Ok... thanks.
>What I am really looking for is the source of a probibition on using http:
>scheme URI (without fragment Ids) to denote real-world things and abstract

There is none in the current spec.

>Tim uses the strong word 'illegal' above. I continue to fail to
>understand why he feels so strongly that this is illegal.

Er,...  because he wants it to be?  But to be 
more charitable, it does kind of make sense given 
the common useage of things like rdfs:subClassOf, 
since one can say that the bare URI there ought 
to denote the document containing the 
'definition', since such bare URIs are used for 
that purpose in the OWL 'imports' assertions to 
refer to ontologies. In fact, on reflection, that 
might be a better citation, eg see


On the other hand, the semantics of 'imports' is 
tricky to get right and not yet fully 
articulated; and it seems risky to impose a 
condition which depends on such a shaky 
foundation as fragIds and which already has 
large-scale deployed code in violation.  My own 
view is that there really is no real *need* for 
this kind of a ruling, and that to try to impose 
it on the entire Web is like, to use a fine old 
expression, pissing into the wind.

BTW, a weaker ruling could be that any 'bare' 
HTTP URI ending in ".owl" should be taken to 
denote an OWL ontology. That is sufficient to 
carry the burden of OWL importing while freeing 
up other URIs to denote whatever anyone wants 
them to. Or, we could just say that the semantics 
of 'owl:imports' is such that if you try to 
import anything that is not an OWL ontology then 
nothing happens, which makes it a purely OWL 
internal matter, which seems tidier to me.

>His appeal to
>"what the RDF language spec says." appeared to be in support of the claim of
>illegality. What the RDF spec says is that URI refs with frag ids CAN be
>used for this purpose (denoting the thing described by some fragment in an
>RDF document) BUT it does not 'outlaw' the use of URI refs without frag ids
>(and particlarly http: URI without frag ids) for the same purpose - as in
>your original RDF fragment, or the 2nd fragment above which 'illegally' uses
>an unfragmented URI to denote an RDF class.


>This is the substance of httpRange-14 [1]. I think that this is tangled up
>in the distinction this thread has been discussing of an information
>resource (ie. a source of representations) and the thing the representation
>is some sense about (ie. the weather in Oaxaca or a 'galaxy far far away').
>BTW had you consider the effects of placing a super-massive black hole
>'on-the-web' or at least in close proximity to it :-)

Yes. Hey, for all we know we may already be inside an event horizon.

>[1] http://www.w3.org/2001/tag/ilist#httpRange-14
>>  >>I was under there impression that RDF gave URI References (2396
>  > >>terminology) an entirely opaque treatment.
>>  The semantics treats them completely opaquely: however, existing
>>  RDF/DAML/OWL *practice* seems to often follow the convention outlined
>>  by Tim, where a URIref of the form ex:place#thing  denotes an entity
>>  described by some RDF which can be found at the address ex:place.
>A common practice or convention, yes, but what makes it so essential as to
>regard other practices/conventions as illegal?

Im probably the wrong person to ask :-)


IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes
Received on Monday, 21 July 2003 18:26:00 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:00 UTC