[open-bibliography] OKFN blog: Bibliographica, an Introduction

Fri May 28 13:58:34 UTC 2010

Jo Walsh <jo at frot.org> wrote:

> Wrote some notes on the VIAF project mentioned in the last email - 
> hadn't come across before but appears to have been around awhile -
> OCLC effort to integrate authority files from large national libraries,
> some custom RDF vocabulary work, re-use of SKOS and FOAF.
> http://doingdata.posterous.com/on-viaf-and-other-recipes

Extracted from the above url:

A service like VIAF could be created by aggregation; each library coins
its own URI namespace to hold its own authority records. Implicitly,
because the record is within the library's domain name, the
authorityAgency is the holder of the domain name. Bring the records
together by aggregation, and then run the OCLC name-similarity and
linking algorithms over the whole collection, while providing anyone
else with the means to do the same. 

This is the purpose of the BKN project  http://people.bibkn.org/
currently with other sources of name authority data besides what libraries could provide.
A simple record format BibJSON http://www.bibkn.org/bibjson/index.html
allows any agent to express what they know about a name and its connections.
Datasets in this format can then be loaded to the aggregator which converts to RDF.
We have yet to implement name-similarity algorithms over the whole collection, but
we are doing this over subsets of the data.
We are in the process of moving this project to an OKFN server, and looking for contributors of data, 
code and expertise to this effort.
Pointers to open source name-similarity and linking algorithms would be appreciated.

--Jim

----------------------------------------------
Jim Pitman
Director, Bibliographic Knowledge Network Project
http://www.bibkn.org/

Professor of Statistics and Mathematics
University of California
367 Evans Hall # 3860
Berkeley, CA 94720-3860

ph: 510-642-9970  fax: 510-642-7892
e-mail: pitman at stat.berkeley.edu
URL: http://www.stat.berkeley.edu/users/pitman