Identifying Persons
My lunch with Jim Pitman got me interested in looking at bit harder at the issue of identifying people and organizations in the metadata. This can be done authoratively or in an non-authorative distributed way. An example, how do I specify which "Jane Doe" I mean I use her name? We are asking this question specifically in the context of browsing and searching the electronic academic literature.
I thought immediately of two possibly relevant specifications, specifically FOAF and MADS. I have been spending time this afternoon looking at both.
MADS and the Library of Congress Authority system and the OCLC service
The Library of Congress has developed an elaborate catalogues of names and subject headings in its cataloguing practices. A place to start learning about it is ![[WWW]](/www/moin/htdocs/classic/img/moin-www.png){kind=small}
Library of Congress Authorities (Search for Name, Subject, Title and Name/Title):
-
Using Library of Congress Authorities, you can browse and display authority headings for Subject, Name, Title and Name/Title combinations. You can also download authority records in MARC format for use in a local library system. This service is offered free of charge.
I used the search page to do a "Name Authorities Heading" search on "Czeslaw, Milosz" to find out that the corresponding Library of Congress Control Number for Milosz is 50033350. (Do the search yourself to confirm this fact. Because of time-dependent sessions, I can't provide an easy way to link to the right LoC record here.)
Alternatively, you can use the LC Name Authority File from OCLC:
-
OCLC Research's LC Name Authority Service provides a way for systems to check names in their databases against the Library of Congress Authority File.
which turns out to be much more functional. If you go to the search form and perform the name search for "Milosz, Czeslaw", you can get the name authority record in either HTML or MARCXML. This service is queryable as a SOAP service (see the corresponding WSDL file.) Neat.
Looking at the name authority records for Milosz and for Bach, Johann Sebastian, 1685-1750 (HTML) was really eye-opening for me. I really didn't understand the huge number of variant spellings for Milosz -- and even more for Bach.
Among the variants for Milosz are:
-
Miłosz, Czesław
-
Milosh, Cheslav
-
Milošas, Česlavas
-
Kózka, B. B. (???)
-
Miloš, Česlav
and for J. S. Bach:
-
Bach, Johann Sebastian
-
Bakh, Iogann Sebastian
-
Bakh, Y. S.
-
Bach, Jean Sébastien
-
Bach, G. S
-
Bach, Jan Sebastian
-
Bakh, Ĭokhan Sebastian
-
Bakh, Yohan Sebasṭyan
-
Bach, Juan S.
-
Bach, Giovanni Sebastiano
Hmmm....I think that I will need help from a MARC expert to interpret what the record really means -- but I take them as listing variants in the names.
At any rate, the Library of Congress through its naming authority does author a central catalog of names, an indentifier for the world of authors and creators in the intellectual and scholarly world. It's wonderful -- but it's useful only if your work happened to have been catalogued by the Library of Congress (at least, that's my understanding).
Can we use the Library of Congress system as a jumping off point? I can't imagine that LC would get in the business of making official records for everyone in the scholarly/intellectual/artistic community. Maybe we can use the conceptual framework that LC has created. That's where I began looking to MADS as a model for how to name authorities (persons and other entities).
FOAF
People who are trying to build infrastructure for the online communities represented by such things as bloggers and social software systems have been looking into how to represent personal information and social connections in a easily machine parsable way. The the friend of a friend (foaf) project (FOAF) is "about creating a Web of machine-readable homepages describing people, the links between them and the things they create and do."
To give myself a feel for FOAF, I used FOAF-a-matic -- Describe yourself in RDF ("a simple Javascript application that allows you to create a FOAF ("Friend-of-A-Friend") description of yourself") to build a FOAF file for myself: MyFOAF.
FOAF and MADS?
I have a feeling that FOAF and MADS are playing in a similar, overlapping, but distinct space -- though I might be totally off here. I will want to learn more about the two as I move forward. A specific motivation is making it very clear that when I'm talking about CzeslawMilosz that I'm talking about the person identified by the Library of Congress in its authority file as HTML. For such a prominent figure as Milosz, it might be overkill to go to such lengths as tying all this information together. But for all the various "John Smiths" out there, it might be helpful to use such methods as name authorities to disambiguate the various agents out there.
The accession number for identifying myself in Melvyl
Since I've never written a book, my name is not in the LC Name Authority system. But I did write a doctoral dissertation, which is filed in the library on campus. I discovered that Melvyl provides an "accession sequence number" (which for me is 079674955) to pull up what I wrote:
I thought that accession sequence number had some sort of unique identifier for an author until I went to did a search for J. S. Bach by accession number and saw that there were other numbers also representing J. S. Bach. Why more than one number? Hmmmm.
