Ideas from Workshop on National Strategy for Shareable Local Name Authorities

Earlier this week I attended a two-day workshop on National Strategy for Shareable Local Name Authorities. The meeting came together as a result of an IMLS National Leadership Grant to Cornell University. From the grant abstract:

Libraries create local authorities to serve a variety of purposes, usually within an institutional context; but these authorities have significant potential for reuse at other cultural heritage organizations and beyond. The April 2015 IMLS National Digital Platform Forum report emphasized the importance of enabling technologies (e.g. interoperability via linked data) and radical collaborations in supporting the mission of the cultural heritage sector. By facilitating a national forum, we plan to identify solutions for facilitating the creation of more shareable authorities. This is a critical area in which a highly collaborative effort can promote interoperable metadata and linked data readiness.

Controlled vocabularies are critical to effective metadata…it is how we bring together the works of a creator as well as works on a particular topic. A phrase that was attributed to Anna Neatrour from the University of Utah stuck out for me: authority control without knowing they’re doing authority control. How is it that we could create systems that allow the user to select or create good authority records while minimizing what they need to know about authority control. Sarah Ross from Cornell pointed me to the OCLC searchFAST user interface as an example of how this can be done:

http://fast.oclc.org/searchfast/?&limit=keywords&facet=all&query=history&sort=usage+desc&start=0#&single=fst01411628&fullview=marc&sep=dagger

“Effective” is partially in the user experience for selecting records from controlled vocabularies, as the searchFAST interface demonstrates. Another part of “effective” is having a good source of controlled vocabulary values to pick from. Libraries and cultural heritage organizations are chock full of formally curated vocabularies (thinking LCSH, LC Name Authority File, the VIAF aggregation of national name vocabularies, etc) and a whole bunch of other, locally-produced authorities. This workshop was about finding ways to make the locally-produced authorities more visible to others.

One thought was to have a list or registry of locally produced name authority files that can be machine processed. That way an aggregation agent could bring them together in a system that could be used for authority control processing. (Perhaps to be called Reconciliation-as-a-Service.) There would be some work, of course, to define data elements in the aggregation that an operator would use to distinguish between many possible versions of a name. That is where a good user interface would be critical for efficiently and effectively selecting the correct name, or determining that no correct name exists and a new name identifier needed to be created.

There is a group of people from the workshop that are interested in working on this problem. I think there are aspects of the problem that would make for useful apps in FOLIO systems. More thoughts welcome!

I think that’s all quite correct. On the UI side, how to make it efficient and pleasant for people to use controlled vocabularies. On the data side, how to have local controlled vocabularies (perhaps especially names) to augment when commonly-shared vocabularies (e.g. LC Name Authorities) do not have an entry that supports a local need. And then later to establish equivalencies as the commonly-shared vocabularies grow to include suitable terms.

Having such a facility for authorities in FOLIO will be essential, and it makes sense to structure it in a way that other modules can tie in to it. I can imagine a future where, for example, the ILS functions, institutional repositories, digital collections and more live on FOLIO and will all need this such a facility.

[As an aside (and perhaps a bit out there): what does such an authorities facility have in common structurally with, say, the identity infrastructure? Are there useful parallels and essential differences that come into play when moving on to the design and engineering of these features? I.e. is there anything that one functional area can learn from the other?]

We are seeing things the same way, Tod. I wanted to augment one point you made:

Rather than “live on FOLIO” I would envision a future where specialized systems (such as ArchivesSpace for EAD and Hydra/Islandora for digital collections) would make use of programming interfaces provided by FOLIO to learn about and manage controlled vocabularies. In other words, those specialized systems would call upon FOLIO to manage local names so that those names could be shared across many systems.

You bring up an interesting point about identity systems related to name authorities. I guess that is what ORCID is doing for us, no? I hadn’t thought about parallels there, though. What did come to mind is the distributed nature of the Domain Name System (DNS) and if there are parallels there.

I would support that augment! A re-usable authority service on the platform is more attractive that a walled module. We’re definitely talking reusable systems and breaking down of silos.

There are two immediate parallels with DNS:

1. the distributed nature involves a trust chain for the statements involved, and
2. caching (interpret this broadly) must be effective for the system to work at any scale.