As the scope of library collections and descriptive efforts has grown, existing methods of authority control have shown strain. Many libraries and cooperative initiatives have been experimenting with new methods of managing information around people associated with creative works. This talk will explore the reasons for these moves, and their implications for technical services workflows, library cooperation, and discoverability.
17. • Scope – high % are monograph authors
• Model – flat/unclear (pseudonyms, titles, names, etc etc)
• Mandate – only when and to the extent a name form
needs to be unique
Intellectual scope, model, mandate
18. • Applicability to special collections / non-published
materials is weak
• Tied with historical decisions to not catalog articles
Authority control scope – material type
19. • Models differ
• Quality control differs
• Duplication
• Different “clumping”
• Name forms
Authority control and scale
23. 600 1 7 ‡a Vollmann, William T.
‡2 fast ‡0 (OCoLC)fst0018706
Becomes…
600 1 7 ‡a Vollmann, William T.
‡2 fast ‡0 http://id.worldcat.org/fast/187068
“Linkification”
25. <http://www.worldcat.org/oclc/25509331>
# An Afghanistan picture show, or, How I saved the world
a schema:CreativeWork, schema:Book ;
schema:creator <http://worldcat.org/entity/person/id/2639915820> ;
# William T. Vollmann
library:oclcnum "25509331" ;
library:placeOfPublication <http://id.loc.gov/vocabulary/countries/nyu> ;
library:placeOfPublication <http://dbpedia.org/resource/New_York_City> ;
# New York
schema:exampleOfWork <http://worldcat.org/entity/work/id/200636202> ;
# Afghanistan picture show
schema:genre "History"@en ;
schema:genre "Personal narratives"@en ;
“Entification”
26. • Converting data is cheaper and easier
• Garbage in, garbage out
• Consistency is good for sharing
• Some work is “computationally expensive”
Limits, investments, and tradeoff
28. WorldCat Person Entity Lookup Pilot
• How does it work?
– Users can find and identify persons with common IDs
– Users can enter text (for example, a personal name)
– Information returned includes other IDs, birth/death dates,
roles, etc.
• Pilot participants
– Biblioteka Narodowa (Poland), Cornell, Deutsche
Nationalbibliothek (Germany), Harvard, LC, NLM,
Stanford, UC Davis, Pepperdine, swissbib, Drexel
36. Lessons learned
• Discovery potential
• Potential for wider contributor base
• Need to include more non-authority data
37.
38.
39.
40. Lessons learned
• Discovery potential
• Potential for wider contributor base
• Need to include more non-authority data
• One-by-one is good for resolution, but not for conversion
41. Lessons learned
• Discovery potential
• Potential for wider contributor base
• Need to include more non-authority data
• One-by-one is good for resolution, but not for conversion
• Linking is fun
44. • Improve recall by controlling as many names as possible
• Improve precision by improving authorities
• Participate in your friendly local NACO funnel
• Add MARC records to WorldCat for special collections
…to make discovery work better?
45. • Consider adding local identifiers as stopgap
• Document and share local headings
…to make metadata workflows better?
46. • Surface hidden collections by focusing on identifying
personal names and organizations
• Use WorldCat, ArchiveGrid, etc to find other institutions
with similar collections (or authors?)
… to expand the scope of description?
Name forms are created only with enough detail to differentiate. Usually a year of birth is enough, but see the example here where even a month was not specific enough.
Even with this much detail, a library will still have to deal with metadata that mentions “John Chapman” without any more info.
At OCLC, we knew we could help libraries identify and pull together resources in a more comprehensive way – so we created Persons as a way to do this.
Name forms are created only with enough detail to differentiate. Usually a year of birth is enough, but see the example here where even a month was not specific enough.
Even with this much detail, a library will still have to deal with metadata that mentions “John Chapman” without any more info.
At OCLC, we knew we could help libraries identify and pull together resources in a more comprehensive way – so we created Persons as a way to do this.
Name forms are created only with enough detail to differentiate. Usually a year of birth is enough, but see the example here where even a month was not specific enough.
Even with this much detail, a library will still have to deal with metadata that mentions “John Chapman” without any more info.
At OCLC, we knew we could help libraries identify and pull together resources in a more comprehensive way – so we created Persons as a way to do this.
First phase was ID-based, second phase allowed for text-based searching
This is a view of a prototype for displaying Person Entities. Note that we also pull in information from other vocabularies to build this view.
We use a numebr of sources to show additional information about the person. This is crucial for disambiguation between similar names (the “John Smith” problem)
We capture alternate name forms, and these are indexed so users can get to the entity through a number of alphabets
Also important are “same as” relationships, which allow users to disambiguate and to get more information. It’s also useful to pull together displays such as this one.
In many cases we can even show family relationships.
Libraries need to serve their communities, and in doing so, local / special materials are important.As we expand beyond traditional authority files, we open up new opportunities and challenges for managing names and other information about persons.