-----BEGIN PGP SIGNED MESSAGE-----
> You seem to have missed part of the discussion about identifiers. I point you to:
> The thread begins here, but unfortunately the archive is not working correctly
> and some posts (esp. those from LC) do not display:
Maybe I have been following that subthread to lazily, but I
think the discussion there mainly was about what URIs could
be used to identify an instance of a bf:Identifier (or comparable
constructs in any scheme) as a resource and the advice was
to treat them as anonymous nodes whenever possible, or at
least and completely counter-intuitive to provide them with
almost meaningless, as local as possible, URIs.
Now a bf:Identifier in the examples has one identifierValue but
I think there should be no restriction (e.g. it should be possible
to list an ISBN-10 and the corresponding ISBN-13 within the same
bf:Identifier construct). If now I not only happen to know the
ISBN-as-string of my resource but also the URN:ISBN URI then I
also need to express the fact that these actually mean the same
within the semantic context of "the ISBN system".
Listing strings and URIs (perhaps not as identifierValue but
as an hypothetic identifierURI) within the same container is
not contradictory on a formal level, I think. If we now assume
that the bf:Identifier is somehow bound to a resource (e.g.
when it is contained within an RDF:Description element) then
the semantics of a bf:identifierURI would be comparable to
that of the string-valued identifierValue: Whenever a resource
contains a bf:Identifier listing this URI, the two resources
shall be considered the same.
Considering bf:Identifiers listing multiple values reflects the
concept of "abstract identifierd" expressed in my previous post.
Maybe that is not the intended usage, because it would not
allow to qualify the individual string representations. If one
would like to state "this form ist the ISBN-10" and "that
form ist the ISBN-13" one would need two different bf:Identifier
constructs and the problem arises to provide statements expressing
their equivalence, which also in the situation of string
identifiers rises the problem to identify these with URIs
void of any ISBN meaning.
> And one of those from LC, 7/11/14, from Ray Denenberg, states:
> "·I believe it has been clearly demonstrated by this discussion that a URI
> should not be one of the “identifier schemes” for bf:Identifier."
> And in the thread that begins on 7/10/14 with a post by Karen Smith-Yoshimura, I
> believe that we demonstrate that using as subject a URI from a third part does
> NOT imply that the statement was made by that party. This is one of the
> fundamental "truths" of the semantic web - that anyone can say anything about
> anything (AAA), and the URI does NOT indicate provenance of the statement (triple).
Thanks for the pointer, this was the thread I was trying to remember in
my post. Kevin Ford especially raised some warnings about "nasty things
that could happen" but at the end of the thread the optimism that suitable
mechanisms for denoting provenance of graphs will be available dominated
the result of the thread. IIRC I merely mentioned fears of some unspecific
"graph pollution". Beginning to think about it much of the decision of
using "foreign" URIs vs. crafting my own depends on the nature of the
resources and my own knowledge about them (how sure can I be that my
statements are really about the resource you had in mind?), my context
(I should re-use those URIs already employed by my institution or my
favorite community) or my intended audience (when my vocabulary
differs I at least will provide the URIs you already happen to know)
and probably more. And already these three points conflict with each
other. Therefore our traditional identifiers will even in semantic
web contexts serve their purpose of faciliating cross-community identifcation...
Thus maybe I should only warn that we cannot rely on the other extreme:
Even bf:instance s are something other communities would consider to be
classes and therefore there are no real obstacles that two agents indeed
use the same URI for identifying a resource, they still do this independent
of each other and they /could/ have decided otherwise. Thus for a given
resource there always will be many URIs in use and the set of all statements
formulated with bibframe vocabulary will not be a substitute for a FRBR
manifestation level kind of authority file.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
-----END PGP SIGNATURE-----