My 2017 ALA Annual Conference experience got off to a bit of a bumpy start, because my flight was delayed about  two hours and I wound up missing a dinner meeting with the editorial board of Technical Services Quarterly. But overall, it was more productive and enjoyable than my 2017 Midwinter experience, much of which I spent sick as a dog. My main takeaways had to do with the transition from MARC to BIBFRAME and the relationship of identity management to authority control.

Both of these things involve URIs. Now, you may already know what URIs are, but if not, I’ll try to give a brief explanation of them and how they fit into the world of cataloging and linked data. URI stands for Uniform Resource Identifier, which means it is a string of characters that link to an internet resource. A URL (which we’re all familiar with) is a type of URI. Linked data in cataloging means using URIs in place of text strings that need to be updated manually (or by batch processes). It’s sort of like the way we use location codes in the back end of the Voyager catalog and how they display to the public in VuFind. When we moved the microfilm collection from the 4th floor to the 1st floor, we didn’t go into every record for every microfilm title in Voyager and change them from 4th floor to 1st. The Tech Team guys changed a single value in a table, so that the Voyager location code “MIC” was now understood by VuFind to mean 1st floor instead of 4th. You can think of the Voyager location code as being somewhat like a URI. Eventually, our bibliographic records will consist of a series of URIs, one for the author, one for the title, one or more for the subject(s), etc. The URI code will remain constant, but the meaning attached to that code can change and be updated as necessary at a central URI registry. So, when, sometime in the URI-based cataloging future, Stephen King dies (assuming he *can* die, and isn’t in fact undead) and his death date needs to be added to his authorized name entry, there will simply be a single change made in a registry for King’s URI, and all library catalogs across the world that have that URI will display the proper death date. That’s linked data (or at least, it’s a component of what linked data is). This will be much easier than the current system where a change has to be made to the authority record and that change needs to be made in catalogs through local updating processes, whether manual or automated.

A number of sessions that I attended at ALA recommended that catalogers should begin putting URIs into MARC records, so that we can begin moving toward a linked data environment. The MARC format is supposed to be replaced by BIBFRAME, but BIBFRAME is still under development. The thinking is that we will move to a hybrid environment where our MARC records have some features of BIBFRAME records. This would parallel how we have moved to a hybrid RDA environment, where our records that were originally coded according to AACR2 rules have some RDA coding in them, with the expectation that we will eventually move to full RDA-compliance.

One place where I heard about the idea of hybrid MARC/BIBFRAME records was the Library of Congress BIBFRAME Update Forum. Beacher Wiggins and Sally McCallum told the crowd that LC is conducting a second pilot project on BIBFRAME, which involves experimenting with different types of material and languages, as well as developing training materials. They have also converted the entire LC catalog of 17 million MARC records to an experimental database of BIBFRAME records to catalog against. These converted BIBFRAME records need an amount of fine tuning and human intervention, and LC is examining these records with an eye toward fine-tuning the BIBFRAME conversion process. The bottom line statement they made that really stood out to me was that we will be working with MARC for the foreseeable future. Which means that hybrid MARC/BF records are probably inevitable.

Another interesting presentation I attended was “Identity Management or Authority Control?” by Jennifer Liss of Indiana University, which she gave at the Cataloging Norms Interest Group. Liss talked about an initiative by the Program for Cooperative Cataloging (or PCC) to develop a means for more libraries to participate in the identity management process. She began by pointing out that authority control is expensive, and that only 15% of U.S. libraries can add records to the Library of Congress Authority File as members of NACO (which is the Name Authority Cooperative Program of the PCC). NACO has high training requirements, which are expensive, and high work requirements (even small member libraries are expected to contribute 100 new name authorities per year). PCC was tasked with developing a NACO Lite program that would lower the barriers for NACO participation. This program would emphasize identity management, not authority control. Authority control for a person is about making an authorized access point, which is a particular text string associated with a person. Identity management would involve creating a URI for a person. Yes, at least one text string would be associated with the person (their name, or at least one version of it), but multiple text strings could be associated with the URI. And if and when a NACO member-library got around to it, they could create an authority record for the name. This would allow small libraries to create URIs for local authors who have no associated LC authority record. Furthermore, the NACO Lite program is looking at ways to put these URIs into MARC records. The project is still in development, but it seems like a good, common-sense development that would get more libraries involved in the process of managing data about identities.

That’s probably enough (if not more than enough) about my ALA experience, but I did want to mention, for folks who have been at ZSR for a long time, that I ran into Jim Galbraith in the convention center. He’s still Director of the Library at the Corning Glass Museum, and he’s engaged, which is a nice development.