I attended the recent bibframe talk given by Kevin Ford at the NLNZ where in the question time and discussion afterwards we had some robust conversations which were limited by the time available.

What it seemed to boil down to was my perception that all of the bibframe work I'd seen centered on monograph-like works and pre-digital methods of acquiring and holding bibliographic works rather than the much more complex kinds of works, acquisition and holding currently in use.

Kevin's suggestion was that I post to the list with my concerns. Thus this post, which attempts to capture some of the things I've not seen in bibframe (which doesn't necessarily mean that they haven't been done, I'll admit I've not been following the list as I might recently). It reflects only me and my thoughts, not my institution.


Journals

As an academic library, we purchase many academic journals as packages. Each package is associated with a license. Each package contains one or more date-ranges of one or more journals. Each journal may be digitised for some date ranges and born-digital for others. Each journal date range consists of a number of issues. Each issue consists of a number of articles. Journals which are peer-reviewed contain at least some peer-reviewed articles, but may also contain bibliographically and academically important non-peer reviewed articles (i.e. http://www.nature.com/nature/journal/v485/n7396/full/485041e.html ) and retractions. Many journals also expose internal metadata (dates of submission, etc).

Packages are purchased from vendors and held on platforms. Vendors may subsume one another. Platforms may merge, split or collapse. Platforms may vanish, systematically breaking links (as http://www.metapress.com/ appears to have done). For some date ranges of some journals we have third party preservation arrangements (http://www.portico.org/) Platforms may require access mechanisms (IP addresses, complex proxy stanzas (https://www.oclc.org/support/services/ezproxy/documentation/db.en.html)). Content on some platforms is available end-to-end over privacy-preserving HTTPS connections, other platforms use some HTTPS but HTTP for doi resolution, some use pure HTTP. Platforms may or may not have various interoperable methods for statistics gathering (http://www.niso.org/workrooms/sushi etc).

Platforms have various policies and capabilities for having their content exposed to and reused by third parties and tools acting on our behalf, for example discovery systems to harvest full text, citation counters and journal rankers, etc.

That's a stack of complexity, particularly considering I've left out some purely internal dimensions (funding streams, institutional restructuring, audit trails, exchange rate tracking, etc).


Patron-driven acquisition

Patron-driven acquisition involves us surfacing digital titles we don't own as search results to our users and then purchasing those titles when a user makes a certain amount of use of the title (first access, first N pages, etc, etc). Patron-driven acquisition is important because it breaks the implicit assumption that records in our library catalog are the records of 'our' holdings. Due to patron driven acquisition, we handle and ingest many more bibliographic items that we are likely to 'hold' in a traditional sense.

Certain vendors are giving presentations which portray this as an existential crisis for "The Catalog". While they are of course completely, utterly wrong, initiatives such as bibframe silently ignoring things such as PDA isn't helping dampen the flames of speculation.

 
Authority control

Authority control for people is has been discussed for bibframe (mainly VIAF and orcid.org). Authority control for is broader than that, however, and almost certainly needs to be done for serials and probably individual works.

https://scholarlyoa.com/other-pages/hijacked-journals/ is a list of apparent attempts to profit (directly or indirectly) from the current confusion in naming academic journals. Proper authority control is the only solution I can see to this.

It's also not clear to me how to separate complex works, For example the consider the multiple works known as "The hitchhikers guide to the galaxy" or "The Tales of Beedle the Bard". This is related to an ongoing issue we have in our discovery system: where a (book/film) review and the work being reviewed have the same name, endless confusion ensues. I'm assuming that authority control would be used to do this kind of disambiguation, but a concrete example would be very useful.

I hope this helps.

cheers
stuart
--
I have a new phone number: 04 463 5692
https://www.facebook.com/VUWLibrary / https://www.facebook.com/TKMPC