W3C home > Mailing lists > Public > public-lld@w3.org > July 2010

Re: [open-bibliography] MARC Codes for Forms of Musical Composition

From: Young,Jeff (OR) <jyoung@oclc.org>
Date: Wed, 07 Jul 2010 18:40:55 -0400
Message-ID: <00ae01cb1e25$6fe204ab$d71dae84@oa.oclc.org>
To: "Erik Hetzner" <egh@e6h.org>, "public-lld" <public-lld@w3.org>
Assuming we aren't being arbitrary, a lightweight way to keep the
individuals separate is to use a hash URI convention like so:

http://example.org/foo/12345/#foaf:Person, #skos:Concept, etc.

Jeff

Erik Hetzner <egh@e6h.org> wrote:

At Wed, 7 Jul 2010 16:53:24 -0400,
Houghton,Andrew wrote:
> > If resources are constrained to having one rdf:type, then we are going
> > to create many new, unnecessary URIs, as Ross says above, breaking the
> > web architecture best practice:
> >
> >   A URI owner SHOULD NOT associate arbitrarily different URIs with the
> >   same resource. [1]
> >
> > 1. http://www.w3.org/TR/webarch/#uri-aliases
>
> This is a gross overstatement of the Web Architecture best practice...

Hi Andy,

You clipped the part where I addressed this:

> > [Ö] you can use two identified resources (URIs) to keep track of two
> > different concepts (e.g., a person as a person, or a person as a
> > concept).

Iím not arguing against using different URIs to describe different
resources. I was trying to argue against an arbitrary constraint or
best practice that any given resource can only have one rdf:type. This
constraint would force, as in Rossí example, a user to create
associate multiple URIs with the same ISBN, because there are many
different ontologies that have a book type.

While it is useful for us not to come up with arbitrary new RDF types
that mean the same thing, or have overlapping meanings, it is the
nature of the beast and we have to deal with it. Deciding that a
resource has one & only one type (or rather is a member of only one
type hierarchy) is not helpful, because it will lead to URI
proliferation for any resource which should have multiple types,
perhaps because we want to use support two ontologies which have types
with overlapping meaning.

> For example, a foaf:Person and a skos:Concept are *different*
> resources. In many ways they behave at the Web Architecture level as
> variant representations of a generic resource. In linked data terms
> the Real-World Object for the conceptual thing has a URI and the Web
> Documents for the foaf:Person and skos:Concept variant
> representations have separate URIs. The Web Document URIs *are not*
> URI aliases, they are separate resources unto themselves, as
> described by the TAG GenericResources-53 decision or the TAG finding
> that came out of the decision [1]. A URI alias, as described by your
> reference applies to creating multiple URIs, for example, to the Web
> Document describing the skos:Concept variant representation. My
> prior message clearly shows that I wasn't creating URI aliases for
> either the foaf:Person or skos:Concept and the analogy to the TAG
> GenericResources-53 decision insures that individuals are not
> conflated at the ontological level just as you shouldn't conflate
> Web Documents at the Web Architecture level.

Largely agreed. Iím sorry that I wasnít clear.

best, Erik
Received on Wednesday, 7 July 2010 22:41:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 7 July 2010 22:41:32 GMT