- From: Joshua Allen <joshuaa@microsoft.com>
- Date: Tue, 6 Aug 2002 20:09:49 -0700
- To: "Seth Russell" <seth@robustai.net>, "Peter Bruhn Andersen" <bruhn.andersen@get2net.dk>, "RDF-interest" <www-rdf-interest@w3.org>
But of course a unique URI is the most authoritative daml:UnambiguousProperty... Although, I am curious about why the aversion to e-mail. If the person has no e-mail, and wants to get a URI, the easiest way to get a URI seems to just get an e-mail address. Beyond that, I would use a one-way hash "fingerprint" that was based on some privileged (to the person) information, but which did not reveal that information. For example, a person's PGP fingerprint would work, or any other 128-bit string based on a public/private key pair registered with an authority like Verisign who can vouch for identity. -----Original Message----- From: Seth Russell [mailto:seth@robustai.net] Sent: Tue 8/6/2002 1:23 PM To: Peter Bruhn Andersen; RDF-interest Cc: Subject: Re: Personal URI? From: "Peter Bruhn Andersen" <bruhn.andersen@get2net.dk> > "Can you create an URN for that person? Possibly with a new URN space?" I don't think that creating a URI (any type of uri) is the best solution. Reason is that different entities who want to describe this person will tend to create different URI for him or her. RDF is all about our data being interoperable. See TimBl's graphs below .. switch back and forth between arcs-1 and arcs-3 and see how the data from different sources merges together. But if you call this person <http://foo/bar#Jim> an his school calls him <http://whatever.edu/students#Jim>, then the RDF systems will think you and his school are talking about *different* people. Me thinks it is better to use bNodes and any daml:UnambiguousProperty or just describe the person as best you can: See example below: language: semenglish [ type Person; worksFor "Organization 1", "Organization 2"; fullName "John Joseph Jones". ] Graphs from TimBl <http://robustai.net/mentography/arcs-1.gif> <http://robustai.net/mentography/arcs-3.gif> Foaf: <http://xmlns.com/foaf/0.1/> Seth Russell
Received on Tuesday, 6 August 2002 23:10:10 UTC