Re: [semanticweb] Purl.org offline?

Greetings.

On 2008 Oct 16, at 09:33, Reto Bachmann-Gmür wrote:

>> http://meaningof.example/?as-of=2008&uri=http://purl.org/dc/terms/title
>>
> The resolution of httpRange-14 says that 'If an "http" resource  
> responds
> to a GET request with a 2xx response, then the resource identified by
> that URI is an information resource;',

True, but not, I think, relevant.

> assuming that rdf:Property and
> information resource the casino-site operator could make all
> rdf-document using dc:title to collapse as contradictions by placing a
> document at that URI.

No.  The name of the DC title property is <http://purl.org/dc/terms/title 
 >, which has the form of a URI which happens, right now, to be  
retrievable.  However it doesn't really matter where the information  
about that is retrieved from, and in particular it doesn't matter if,  
when you retrieve that URL, you get casino adverts or porn.  The  
'follow your nose' aspect of the semantic web is a best practice, not  
a rule, and failing to observe it simply adds inconvenience.

The whole discussion about decentralised URLs can rapidly become  
rather fevered, and head off into URNs and madness.  But I don't  
believe it's really a practical problem.

The ontologies http://billy123.backwoods-isp.net/foo.rdf and http://www.loc.gov/foo.rdf 
  are, technically, equally vulnerable.  But one of these is  
_actually_ more reliable than the other, not because of technical  
stuff about server farms and DNS round-robins (though those will  
help), but for old-fashioned bureaucratic reasons such as having a  
paper-based process for making sure they renew their DNS registration  
in good time.

Short version: this is a social, not a technical, problem. (and it  
probably isn't much of a problem anyway -- was anyone really  
inconvenienced by purl.org being down?)

Best wishes,

Norman


-- 
Norman Gray  :  http://nxg.me.uk
Dept Physics and Astronomy, University of Leicester

Received on Thursday, 16 October 2008 09:31:10 UTC