-----BEGIN PGP SIGNED MESSAGE-----
Sorry for the late reply,
Am 03.08.2015 um 16:21 schrieb Karen Coyle:
> On 8/3/15 2:46 AM, Thomas Berger wrote:
>> We'll probably have to do both in even more situations than before
>> (links becoming more important while the text form still being
>> indispensable as proof of evidence for many applications),
>> since the two ways of expressing the fact are complimentary
>> and by no means full substitutes of each other. Thus it will be
>> crucial to never lose the connection between the two for a given
>> instance, i.e. having seemingly independent series "statements"
>> and series "links".
> Why? Every "thing" in RDF can have display forms -- what is the series
> statement but a display form of the series information? We should assu
> that everything that has an identifier also has a display for humans.
subsequent posts by Stephen and Mac brought up the multilanguage
issues I originally didn't have in mind.
We already had seen that a "series statement" is of compound nature,
since it indicates the series (a different entity from the one
carrying the statement) and also the "numbering within series"
which best may be seen as some qualifying information pertaining
to the relation itself.
Within an RDF context there are technical obstacles coping with
that - i.e. turning "part of a text" into a "hyperlink" or
accompanying textual data into collation-friendly numbers or
sortable strings: Whenever there is a numbered subseries within
a numbered series and/or parallel numberings the series
statement is quite hermetic and spreading it out to something
machine processable is not straightforward at all. (Thus
perhaps one "series statement" (transcribed as stated on the
resource) may give rise to one or more complex "series link"
RDF statements (as in stated by the cataloguing process)
at leaet indicating the target resource (the "link") and that
part of the numbering which pertains to the series in question.
Series being "analytic" or that kind of "aggregates" which
evolve over time are a problem, too:
* A series as a work is the result of the activity of editors
and publishers (our concepts of "work" are broad enough
to extend well beyond artistic or intellectual production
to all areas of organized human effords)
* The examples of multilanguage series (and - as by now almost
always - the parallel existence of "online" and "print")
indicate that FRBR expressions of series probably make sense
as a concept
* Actual possession of items may be a faint reflection of the
"series" concept on item level, this is probably completely
irrelevant for bibliographic descriptions, however in
museums and archives "collections" (of objects) are a central
concept and "series" in the sense that a certain subset of
the objects in a given (sub-)collection can be spread out
(e.g. shelved) in a linear manner are a technique to achieve
global order from local ordering
However for FRBR manifestaions or BF instances it is not at all
clear what a series might be so it's probably wise to avoid
instances of series and always link to series expressions (FRBR)
or works (BF). Now for a given item we have the transcribed
series statement, which like everything transcribed is a
property within the domain of BF instances / FRBR manifestations.
And we have the actual series link (qualified by the numbering
specific to the specific "series item") which probably "lifts up"
to at least a expression-expression link ("part-of").
So what one could need:
Some RDF container(?) element of BF instance domain, which
can carry the transcribed series statement *and* any number
of "qualified" BF-work to BF-work links which ~should be
understood~ as pertaining to the BF work implied by the
given instance. This sounds rather messy but IMHO exactly is
the value cataloguers usually add: Taking evidence from an
item and abstracting it to instances/manifestations and
even higher levels (BTW: providing a link to an entity for
an author of a given resource and recording the form of her/his
name used on the item also means closely glueing together
data with domain BF-work and BF-instance: Do we really
understand yet what is going on there?).
Alternatively one could have any number of "series links" with
domain BF-work and "source them" with transcribed data. This
could be "moved" then to the BF work graph (quotes because it
not already sitting there is rather a misconception of us) Here
I'm not sure how to express what is taken from what instance
and how often it really is imprinted there...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
-----END PGP SIGNATURE-----