I like the way you've articulated this, and it brought up another thought. In our usual authority practices, we have added citations to the works of an author to 'justify' choices in heading but also to make clear which 'John Smith' we were talking about. In the environment you describe, the links from an author to his/her work will operate to disambiguate in much the same way. This, I think, supports your notion of the blurring of the distinction between authority descriptions and bibliographic descriptions, given that the links, operating in both directions, serve some of the same goals and are not really primary or secondary in the way bib and auth records are seen in our current environment. 

This fits in quite nicely in the argument I've been making about the withering away of authority records that aggregate those names that cannot be distinguished--necessary when name strings needed to be unique but not when string uniqueness is no longer required.

I hope this makes sense!


On Wed, Nov 28, 2012 at 10:33 AM, Robert Maxwell <[log in to unmask]> wrote:
I've been out doing training and so haven't been able to respond to Kevin's question. John Myers's response, however, has articulated well my concern. In the ER structure that I've been hoping we were moving toward, all the entities would exist on a relatively equal basis and we would be engaged in recording information about them (e.g. persons, works, manifestations, etc.) in descriptions, and then those descriptions would be linked together by specified relationships. In such an environment I don't see a need for a separate authority structure--instead the entity descriptions would stand for themselves, they would justify themselves. In our current environment, at least pre-RDA, the authority records were there pretty much simply to record decisions about the form of name that we pasted into a bibliographic record. This won't be necessary in the future environment because we won't (I sincerely hope!) be pasting authorized access points into bib records; instead we will simply be linking (e.g.) a person description to the appropriate other entity description (work, expression, manifestation, item, or for that matter, another person, corporate body, etc.) and the access point for the person won't appear in the description it's linked to. Under RDA I'm seeing the authority records for persons, families, corporate bodies, works, etc., develop into those entity descriptions and I'd expect, when all this migrates to the new structure, the "authority" part of it to wither away, although we'll probably still need to have preferred forms for labeling purposes. But not for purposes of recording a form that we're supposed to copy or paste into another record.


Robert L. Maxwell
Head, Special Collections and Formats Catalog Dept.
6728 Harold B. Lee Library
Brigham Young University
Provo, UT 84602

"We should set an example for all the world, rather than confine ourselves to the course which has been heretofore pursued"--Eliza R. Snow, 1842.

From: Bibliographic Framework Transition Initiative Forum [[log in to unmask]] on behalf of Myers, John F. [[log in to unmask]]
Sent: Tuesday, November 27, 2012 9:18 AM
To: [log in to unmask]
Subject: Re: [BIBFRAME] [RDA-L] BIBFRAME model document announced

The current bibliographic structure is essentially bifurcated between: bibliographic records that describe and support access to resources; and, authority records that authorize and structure the headings/access points employed within the bibliographic records.  The paucity of information recorded and functionality of our authority records betrays their roots as a subordinate structure to support the main (card) catalog.

The Functional Requirements family of models articulate a structure where all of the associated entities in a bibliographic universe exist on a relatively level and agnostic playing field -- the records for works, expressions, manifestations, items, people, families, corporate bodies, concepts, objects, places, events are all of equal importance to the functioning of the overall database.   Recorded elements are distinctly and uniquely associated with each class of entity.  In particular, records for entities formerly addressed by authority records are significantly more robust, holding elements above and beyond the mere heading/access point.  Elements from related entities across the database are gathered and presented to form the display we traditionally associate with a bibliographic record.

The BIBFRAME model, in its overview articulation on p. 8 of the report, appears to sustain the former model, with its references to Instances and Authorities.  The further details and then the articulation in "Serializing the BIBFRAME model" however, show that things are somewhat better aligned with the FR family than the overview indicates.

There are still some concerns though that:
* the distinct nature of the elements associated with the FRBR Work and Expression entities will be muddled in the BIBFRAME Creative Work class;
* the FRBR Item entity will be inadequately articulated or will be subordinated within the BIBFRAME Annotation class;
* a subtle intellectual bias dividing the catalog into the BIBFRAME Instance class and "the rest" will persist.

Having said that, I think the report is an interesting starting point for further conversation  The devil will be in the details, as they say.  So working out implementation will be a critical aspect, that will trump all the labels we apply.


John F. Myers, Catalog Librarian
Schaffer Library, Union College
Schenectady NY 12308

[log in to unmask]

-----Original Message-----
Kevin Ford wrote:

Dear Bob,

Thanks for the early feedback.  I look forward to your thoughts once you've read the document more thoroughly.

I wanted to take the opportunity to comment on or request further information about this:

> Further, report's apparent continuation of a model that continues the
> division of the database into "authority" and "instance" (which I
> gather is more or less the equivalent of bibliographic records, see p.
> 10 of the report) seems extremely backward to me.

I think it possible that Authorities (People, Places, Topics, Organizations) and Instances (as well as Works) could be seen as equal entry points and therefore not "continue the division of the database," but I may be misunderstanding you.

Even as I write/read that statement, I feel like I am not understanding your point.

Can you tell us *how* the proposed model continues the old model with respect to the "division of the database"?



Kevin Ford
Network Development and MARC Standards Office Library of Congress Washington, DC

> -----Original Message-----
> From: Bibliographic Framework Transition Initiative Forum
> [mailto:[log in to unmask]] On Behalf Of Robert Maxwell
> Sent: Friday, November 23, 2012 6:34 PM
> To: [log in to unmask]
> Subject: Re: [BIBFRAME] [RDA-L] BIBFRAME model document announced
> I haven't had a chance to look closely at the document yet, but it
> does disturb me that "a team from Zephira" appears to have, having
> thought about it for a few months, swept away nearly two decades of
> consideration by the best minds in the cataloging profession by
> apparently abandoning the FRBR model, as Mac points out below. I
> realize not everyone agrees with the FRBR model but I should think
> such a step should not happen simply because of a report from a
> consulting group. Sally McCallum said in her announcement that "like
> MARC, [the model] must be able to accommodate any number of content
> models", which is certainly true, but one would think that at least
> one of those content models might be RDA, which was the entire impetus
> for hiring Zephira to come up with a new model for us. Since RDA is
> firmly based on FRBR and DOES include provisions for describing and
> linking to expressions, it does seem inappropriate that the new model
> should not provide for this entity. I have a hard time seeing how this
> model would be any better a fit for RDA than the current MARC model.
> Further, report's apparent continuation of a model that continues the
> division of the database into "authority" and "instance" (which I
> gather is more or less the equivalent of bibliographic records, see p.
> 10 of the report) seems extremely backward to me. In an ER linked data
> database we would have descriptions of the entities linked by
> relationship links.
> Bob
> Robert L. Maxwell
> Special Collections and Ancient Languages Catalog Librarian Genre/Form
> Authorities Librarian
> 6728 Harold B. Lee Library
> Brigham Young University
> Provo, UT 84602
> (801)422-5568