Use cases for local identifiers

That was the topic discussed recently by OCLC Research Library Partners metadata managers, initiated by Chew Chiat Naun of Harvard and John Riemer of the University of California, Los Angeles. In the library world we are familiar with highly centralized models for authority control, as exemplified by NACO and national authority files. However, libraries clearly do create and maintain authorities or name identifier schemes locally. We discussed the use cases for maintaining local authority files, the barriers to contributing local authority records to NACO or national authority files, the trade-offs of minting local identifiers and possible alternative approaches.

The local use cases cited included:

• Registering all researchers on campus;
• Representing entities underrepresented in national authority files such as authors of electronic dissertations and theses, performers, events and campus buildings;
• Recording entities represented in other environments such as digital library projects and institutional repositories;
• Meeting local needs such as local place names and genre terms;
• Reflecting multilingual needs of the community, such as representing entities in Chinese characters rather than in transliterations, especially important for communities who speak Chinese languages other than Mandarin;
• Supporting “housekeeping” tasks such as recording archival collection titles and representing local series classification practice on a copy of the national authority record.

Almost half of all the metadata managers reported that they do not create local identifiers at all, and those that do overwhelmingly did so “for strictly local use”. But “locally authorized” entities may be useful to others who have information about those same entities and should be more widely available. Many would be willing to contribute at least some portion of their local entities to their respective national authority files, but cited significant barriers:

• Authority record creation is too time-consuming and restrictive.
• The volume of all researchers’ names and the entities represented in special collections and archives is too large.
• Not enough staff have the expertise required for authority work.
• The information libraries and archives have for local names is insufficient to meet national authority standards.
• The institution is not a NACO member and has no means to contribute to a national authority file.

The Authority Toolkit to create and modify authority records developed by Gary Strawn of Northwestern University received high praise for making it very easy for anyone to create and maintain authority records, whether a NACO participant or not.

Minting local identifiers for entities has the potential for future algorithmic matching against national and international identifiers. Some skepticism was expressed about whether the information available would be sufficient to support algorithmic matching that establishes “same as” relationships. Early-career academics are encouraged to obtain an ORCID (Open Researcher and Contributor ID) to enable them to aggregate all their scholarly output, including articles and presentations not normally represented in authority files; this requires that they update their ORCID profiles whenever they change institutional affiliations.

We discussed approaches that would encourage sharing the information about entities gathered currently only on the local level:

• Open up NACO to allow editing of existing authority records so non-NACO contributors could add identifiers to the 024 field of the authority record—e.g., ISNI (International Standard Name Identifier) ORCID and VIAF (Virtual International Authority File) identifiers—and death dates, critical for copyright determination. Authority records for article authors could be enriched by their institutions’ library staff by recording disciplines and institutional affiliations to help disambiguate personal names.
• Establish links from authority files to the entities represented in EAC-CPF (Encoded Archival Context: Corporate Bodies, Persons and Families) used by archivists.
• Encourage submission of “provisional” or “preliminary” authority records that others could enrich and establish later.
• Support batch contributions from local authority files to authority hubs.
• Move to a “hybrid” approach that combines matching by machine with human review and correction.

Thinking of the semantic web environment, the OCLC Research Library Partners discussion participants would like to move away from establishing unique text strings to creating identifiers. Rather than the time-consuming task of establishing a unique text string for a personal name, other information could be used to distinguish entities such as the titles of their works. This is especially important in our international environment where the personal name text strings could be written in different scripts. We have a huge pool of legacy entities that do not have identifiers established that we will need to deal with some day. If we don’t find a way to handle the identifiers that libraries are currently establishing locally, we are exacerbating the problem. Let’s try to address the need for identifiers through automation and expanding the scope of authority work.