URIs for People

Hi Norm,

"The FOAF folks tell me that it's bad karma to assign URIs to people."
- http://norman.walsh.name/2003/05/15/nuts

Really? By whom was this advice given out, and where? Whilst it's
certainly true that one shouldn't identify people *directly* by their
mailbox addresses (foaf:Person != contact:Mailbox; cf. [1])--or their
hashes, for that matter--the only possible reason against using HTTP
URI-refs to identify people is the same reason against using HTTP
URI-refs in general--the hash vs. slash debate.

I personally use HTTP URIs with hashes in my FOAF file [2] to identify
people that I foaf:know.

If everybody starts using different URIs for the same people, of
course, then once graphs are merged from the Web and deductions from
them made, one will end up with a lot of duplicated information. DanC
worked on a forget duplicates N3 filter a while ago [3] [4], and in
the FOAF world (or that of DanBri and libby at least...) the process
affectionately termed "smooshing" [5].

However, the problem of duplicates also extends to people identified
by bNodes--when graphs are merged, by default you can't merge the
bNodes unless you have extra information. The duplicate forgetting
algorithms for URIs apply just as much when smooshing bNode'd people,
or people who're identifed with a mix of bNodes and URIs.

So I think that assigning URIs to people is not bad, but reducing the
amount of duplicates by having a standard database of people might be
a neat idea. Hmm. Though, on the other hand, that would introduce
unwanted centrality to the system.

[N.B. CC'd to www-archive just in case, though all this information is
freely available on the Web already. Feel free to trim it.]

[1] http://www.w3.org/DesignIssues/InterpretationProperties
[2] http://purl.org/net/sbp/info
[3] http://www.w3.org/2000/10/swap/test/forgetDups.n3
[4] http://www.w3.org/2000/10/swap/test/sameThing.n3
[5] http://rdfweb.org/2001/01/design/smush

--
Sean B. Palmer, <http://purl.org/net/sbp/>
"phenomicity by the bucketful" - http://miscoranda.com/

Received on Saturday, 31 May 2003 09:36:07 UTC