Print

Print


On Aug 1, 2014, at 5:34 AM, Thomas Berger <[log in to unmask]> wrote:

> I very much appreciate the shift from "Provider" to "Event" but I think this has to be kept quite general since Bibframe cannot prescribe what kinds of event a certain community deems important for specific materials (some examples above are connected to questions of copyright status and although outside of "bibliographic" questions may be of importance for "library operations").

I second this advice heartily, and I'll add that some of those certain communities have already done the work necessary to describe the events they deem important for their materials, and it would be foolish to suppose that Bibframe could displace that work. For example, we can think of preservation activities: archivists and conservators might like to use PREMIS events to discuss that work. Or perhaps folks in broadcast media would like to describe some of the kinds of events Thomas Berger mentioned using EBUCore PublicationEvents. I would like very much to avoid having to try to define and connect the same (real) events involving a given resource over and over in different languages, especially when only some of those languages will be appropriate for the actual resource described.

Maybe there is a general principle available here: if there isn't obvious agreement on the details of how to model a particular kind of description, either don't model it in detail, or if necessary offer detailed modeling as an extension to core Bibframe. It's not possible to "be all things to all people". I would like to enjoy the use of a small, clear, flexible Bibframe model with which I can integrate other models (like PREMIS or EBUCore).

---
A. Soroka
The University of Virginia Library

---
A. Soroka
The University of Virginia Library

On Aug 1, 2014, at 5:34 AM, Thomas Berger <[log in to unmask]> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> 
> 
> Am 01.08.2014 09:36, schrieb [log in to unmask]:
>> I'm not sure how I can understand this. From my non-cataloger perspective,
>> I understand you see a manifestation/instance that is published "once". But
>> this is just a monographic publication. What about others?
>> 
>> There are many events in the lifecycle of a work that describe the process
>> of "putting the work to the public", such as
>> 
> ...
>> 
>> Many of this events are essential to library service systems which have to
>> determine if the work is available (for purchase, for licensing, for
>> cataloging, for reference, for circulation, for inter library loan etc.)
> 
> And when you leave the domain of published text material many more events
> are relevant on different levels:
> 
> - - (first) broadcasting of a TV piece (I remember a regional bibliography
>  in austria which routinely covers these, because some formats include
>  many interviews or portraits of persons of regional importance. Since
>  not only the broadcasting firm but also the specific channel and program
>  are relevant this could often be inherited from a bibliographic
>  part/whole relationship as it might be done for articles in journals
>  or newspapers)
> 
> - - exact submission or approval dates for technical reports (by whom or
>  to whom?)
> 
> - - dates of reception (ingress?) for archival material, date of the post-stamp
>  for letters
> 
> - - date of first public display for works of art (this may well be an
>  exhibition as an identifyable entity, or just "19xx by Gallery yz")
> 
> I very much appreciate the shift from "Provider" to "Event" but I think
> this has to be kept quite general since Bibframe cannot prescribe what
> kinds of event a certain community deems important for specific materials
> (some examples above are connected to questions of copyright status and
> although outside of "bibliographic" questions may be of importance for
> "library operations").
> 
> Any event seems to have dates or date ranges associated with it (from
> years down to the granularity of minutes or seconds), many also explicit
> places (as in cities or towns), and agents (corporate bodies typically).
> 
> Especially for manufacturing or publication events I do not understand
> the concerns about re-use: The manufacture of 1000 copies of book X
> and that of 2000 copies of book Y always are distinct events (even when
> they happen in the same shop at the same time). We prefer to record only
> year granularity and therefore yield identical descriptions for the events,
> re-use however would imply they were the same event - that is not what
> we want to express (I hope).
> 
> And (anonymous) subgraphs to my knowledge are the best and usually
> only way to model a non-trivial data structure like needed for "events"
> in RDF. Trying to flatten that out for some typical library situations
> IMHO just conceals the general mechanics and - as this discussion has
> shown so far - introduces a bunch of new problems.
> 
> viele Gruesse
> Thomas Berger
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1
> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
> 
> iJwEAQECAAYFAlPbX0EACgkQYhMlmJ6W47OdVwQAtRqjW6W2iqMk0bU239xdKRHI
> Byz6t8Urtk5SDHajhMCrgSfJWLOGlnVHm9lbMg7p2ZMdrYAt2i0hmfff9piiLfcg
> YX4wo4TKzmlj8QNZltd5mV8MkYLK7xQlIJ9fvuW4SA0pvk63ZDr6DEMjw/fWlwDS
> obehLn9KwHfR4dK9h5s=
> =5rU3
> -----END PGP SIGNATURE-----