RE: Personal URI?

On 2002-08-08, Danny Ayers uttered to Sampo Syreeni:

>The id numbers mentioned so far have national boundaries - will they
>still work globally?

They will, once qualified by country code and issuance date, just as I
said. After that, they will behave like any other sparse, hierarchically
aggregated namespaces does.

>The vast majority of the world's population don't have access to the
>Internet, let alone email addresses - do we need to id people who will
>*never* be online?

I would say yes, which leads us to question, how do we do that? Mail
addresses won't do, NID's are (sometimes) already here and those people to
which neither of these methods applies, are a problem. (I mean, there are
a number of people out there with Chang as their first name.) But that's
still no reason not to assign NID-based identifiers to those whom the idea
applies to.

>Could you please give me an example that couldn't be equally well served
>using bNodes?

We all know such an example cannot be given -- any URI scheme whatsoever
can be supplanted with a bNode and a suitable daml:unAmbiguousProperty.
The point is, this applies to *everything*. If one uses this construction,
there is no reason to use URI's within the set of RDF nodes. We might as
well ban URIs as subjects and make do with bNodes only.

OTOH, one strong reason to consider URIs for persons is architectures
other than RDF, where concepts such as bNodes aren't available. That isn't
something we would like to consider on-list, but it's still relevant as
far as we're talking about new URI-schemes.

>I can accept that it may be easier to deal with URIs directly, rather
>than having to splay back through bNodes, but is easier to deal with a
>mix of URI-identified and bNode-identified people or bNodes alone?

Precisely. Why permit URIs as RDF subjects at all?

I think the prime benefit of having a NID-based person URN is to be able
to bypass the trouble with merging altogether. In certain contexts, like
within the Finnish society, quite a number of instances already use a
unique identifier, unambiguous against the specific background. Giving
those instances, whatever they may be, a tool which enables the current
infrastructure to be reused would, in my mind, attain huge benefits over
reengineering all of the current software to do RDF-based reasoning just
to cross-reference over datasources. Since NIDs are what is being used
now, NID-based URNs become that very tool.

>I don't believe that's so - a URI represents a fixed point, without
>fixed points there's nothing to reason about.

If we presume there are no literals, you're right. A world where
everything is anonymous won't do. But if we allow either literals or node
labels, we will immediately get the necessary fixed point. It doesn't
matter whether the fixed point is given as a node label or a literal
property hanging off an anonymous node. The identity of a node is any
identifying property it might have. Unambiguous properties are one example
and URIs are another. I might in fact add this to my list of reasons why
literals are Bad. ;)
-- 
Sampo Syreeni, aka decoy - mailto:decoy@iki.fi, tel:+358-50-5756111
student/math+cs/helsinki university, http://www.iki.fi/~decoy/front
openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2

Received on Wednesday, 7 August 2002 19:57:59 UTC