Re: ungetable http URIs

David R. Karger wrote:

>We were playing with some simile data in haystack today and discussed
>briefly the fact that simile is defining URIs in namespace
>http://www.mit.edu/simile/ that are not http-GETtable---ie, they return 404
>not founds.  Apparently there has been some discussion of this
>approach but I don't think I've weighed in.  If we don't ever plan for
>the URIs to be http-GETtable, there's no reason for us to use http
>URIs: we could do urn://www.mit.edu/simile....  This would avoid
>browsers getting confused.  Alternatively (and this is the approach
>I'd like to  see) it would be nice if we had a web server that
>returned some useful RDF when we tried to resolve such URIs---just in
>case the requestor was able to cope with that (as haystack can)
>
>-David
>  
>
I think this is one of the points that Mark objects to in RDF.  Since 
there is no standard processing model for RDF, there isn't a real 
meaning for what a node is, or what it should reference.  As a result 
nodes point to all sorts of non-existant things.   rdf:Descriptions be 
about linking a bunch of resources together into some sort of anonymous 
object, or should the description URI reference a real object:

<> a Person;
      :hasEmailAddress <mailto:kevin@ank.com>;
      :hasEmailAddress <mailto:kevin.smathers@hp.com>;
      vcard:FN "Kevin L. Smathers"
.

-or-

<mailto:kevin@ank.com> a Person;
      :hasEmailAddress <mailto:kevin.smathers@hp.com>;
      vcard:FN "Kevin L. Smathers"
.

That said, I have my own opinions on the matter.  I think that the only 
point of making an identifier be a URL is to assert the namespace 
control that comes with URL's.  If you then say that the URL doesn't 
resolve, and it is thus impossible to tell whether the domain owner 
actually approves of the URL you are holding, then there is no reason to 
use a URL at all.   In my IMS work I've tried very hard to make all of 
the URL's I generate from XSLT point to the resources they identify.  

The only area where I failed to do so (that I know of) is with personal 
identifiers (the 'ocwc' namespace) which I think should point to the 
OCLC person record, but which I've been unable to create an exact 1:1 
mapping for.   The current 'ocwc' identifiers are really just place 
holders for a more permanent mapping of some sort, ideally expressed in 
the MIT name space since MIT is the de-facto authority for names found 
in its own OCW content.

Cheers,
-kls

-- 
========================================================
   Kevin Smathers                kevin.smathers@hp.com    
   Hewlett-Packard               kevin@ank.com            
   Palo Alto Research Lab                                 
   1501 Page Mill Rd.            650-857-4477 work        
   M/S 1135                      650-852-8186 fax         
   Palo Alto, CA 94304           510-247-1031 home        
========================================================
use "Standard::Disclaimer";
carp("This message was printed on 100% recycled bits.");

Received on Tuesday, 18 November 2003 15:51:27 UTC