Re: does RDF require understanding all 82 URI schemes?

David Megginson wrote:

> [snip]

>
>  Lots of people expect that
>  > the best place to find info about
>  >      http://example.org/foo#bar
>  > is, in fact, http://example.org/foo#bar
>
> Who owns the thing being discussed?  Let's say that I want to publish
> information about the Battle of Jutland.  If I use the identifier
>
>   http://www.megginson.com/battles.rdf#jutland
>
> I've left anyone else who wants to describe the same battle with a
> choice between two miserable alternatives:
>
> 1. use the same identifier, and eternally privilege my information
> over anyone else's; or

Wait.  The identifier problem is separate from the description problem.
There can be all kinds of privileged, public, and completely private
descriptions of very public things.  The first problem - a universally
accepted identifer for common "things" - has to be solved with reference
to some common, public repository of identifiers.  This is needed.  There
may be more than one such repository, depending on the authority required
to maintain them.  The best examples of this kind of repository today are
http://www.dmoz.org and wordnet
(http://snowball.ilrt.bris.ac.uk/xmlns.com.html).  Yes there are pitfalls
with publicly maintained knowledge bases, but controlled registries have
practical limitations too...

>
>
> 2. use a different identifier, and lose any easy possibility of
> collating the information.
>
> Yech.
>
>  > > no one's statements
>  > > should be privileged,
>  >
>  > Why not? If you want to know what my favorite color is,
>  > surely an answer from me is privileged over an answer
>  > from, say, someone who hardly knows me.
>
> That's a question that can be answered only on a property-by-property
> basis -- you're a reasonable authority on your favorite colour, but
> the W3C is a better authority on whether you're in its employ, and
> WWW10, on whether you're a speaker.  Subjective statements (Dan
> Connolly is/is not intelligent, good looking, interesting, a good
> coder, etc. etc.)  rightly belong to no single authority, though we'd
> be likely to privilege *any* other authority over the person himself.
>
> In summary, it probably makes sense to treat
> http://www.danconnolly.org/about.rdf as an authority for the specific
> case of Dan's favorite colour, but not for many other things.
> Encouraging resolution of RDF resource identifier URIs is wrong for
> the general case.
>
> All of these problems arise even with something -- a living human
> being -- that has an obvious, legally-enforceable identity.  Many
> (most?) of the things we'll want to describe in a data-based Web --
> ideas, historical people/places/things, etc. -- don't even have that.

There are two things that somewhat ameliorate this problem:

1) You don't absolutely need a unique identifier for something, you can
use an unambiguous reference in some accepted namespace.  Just because
you dont have someone's U.S. Social Security number of an encoding of
their thumbprint doesn't mean you can't disambiguate them.

2) Through the miracle of descriptions, you can hone in on something's
identity.  My name is fairly common, but augmented with my date and place
of birth, I can be uniquely identified.

>

>
>
>  > Er... I'm looking at piles of robust, interoperable software
>  > that uses URIs in this fashion.
>  > http://www.w3.org/RDF/#developers
>  > http://www.w3.org/XML/Schema#Tools
>
> I've written a couple of big chunks of it, and I'm afraid it's neither
> robust nor interoperable, though I've done my best.  The inability to
> round-trip a Namespace URI is the killer.
>
> All the best,
>
> David
>
> --
> David Megginson                 david@megginson.com
>            http://www.megginson.com/

--
Mark Grossman
Geocast Network Systems       Tel: (650)566-3259
190 Independence Dr.          Fax: (650)566-8112
Menlo Park, CA 94025          E-Mail: msg@geocast.com