Re: FRBR and the Web

On Mon, Feb 14, 2011 at 12:58 PM, Nathan <nathan@webr3.org> wrote:
> IR being the class of FRBR Manifestations, Expressions, and Works

I think it would be premature to say this, especially without
reference to a particular IR definition.  For now how about if the M,
E, or W has in some cases special relationship to ... something that
we might want to take to be the referent of the URI.

It's conceivable you could write FRBR records for arbitrary web pages
and/or 'generic resources' by using the ideas of serial, aggregation,
and performance liberally. That seems pretty twisted to me though - it
seems to dilute the spirit of FRBR - and I would think the effort
would have low return on investment.

Avoiding this question is why the post is careful to speak of FRBR
entities being 'cozy' to URIs, not to IRs.

> dereferencable absolute-URIs refer to IRs

That's what httpRange-14 says, but whether it's true or not is a
matter of definition, interpretation, and dispute, ergo going this way
is not likely to lead to any insight.

> representations returned when you dereference an absolute-URI are FRBR Items

Yes, if stored somehow. This is close and is what I had in mind
(sorry, didn't I say that?). One might ask, for example, how many
representations reside on some hard drive. But if the bits coming back
are only consumed as a stream, you would have the bits (a
representation) but no Item (locally stored recording). I don't know
whether such re-performances are covered by FRBR - the communities who
would care about talking about these would be more like radio stations
and copyright clearance centers, not librarians and archivists (the
intended audience of FRBR).

> ?

Received on Monday, 14 February 2011 18:53:10 UTC