Date December, 6th 2011, 16:00 GMT

Channel Meeting was held via Skype and Etherpad

Participants

• Adrian Pohl
• Jim Pitman
• Karen Coyle

Agenda

Report from Germany

#swib11

Adrian reported from the third conference Semantic Web in Libraries (SWIB) which this year took place in Hamburg. See the program (Day 1, Day 2) with abstracts of the talks and presentation slides (video recordings to come).

Talking about the project culturegraph we got into a discussion about matching/deduplication algorithms. It was generally agreed, that the problem in this area is, that much duplication of effort takes place because of best practices and algorithms not being shared widely – neither by public institutions nor – of course – by vendors.

It was discussed which would be the best forum to document best practices and/or collect a list of links to relevant projects: the code4lib wiki? An OKF site? The open wiki of the W3C’s Linked Library Data group?

ACTION: As soon as there is serious work on deduplication at culturegraph.org, Adrian will start a conversation about sharing algorithms and best practices and initiate a forum for doing it.

BVB/KOBV data release

Adrian reported about the release of 23 million records from German library networks BVB and KOBV. (See this post for information on this).
Karen showed interest in a documentation of the MARC-to-RDF conversions.http://epsiplatform.eu/content/bavaria-opens-data-portal. But so far no documentation seems to exist. Jim would like to know how he can get all the data that is relevant to him (concerning mathematics/statistics) out of the MARC/XML or the Linked Open Data. That is a question he generally asks himself when he hears about open bibliographic data releases.

Report on BKN/BibSoup/BibServer/BibJSON progress (Jim)

• BibServer now imports and exports according to documented BibJSON standard.
• Having read Rufus’ answers regarding the questions about thedatahub.org as a data repository (see last meeting’s minutes), Jim is fine with using the Data Hub to upload his open bibliographic datasets.

• ACTION: Exemplar records and working instances for further discussion of entities in BibJSON.

LoC Bibliographic Framework

In the end, there was a short discussion about the Library of Congress’ initiative for a new bibliographic framework based on Linked Data standards. The US-centric as well as library-centric approach was seen as a problem and the need to broaden the approach and seperate activities between agents (LoC for cataloging, W3C for technologies, NISO etc.) was articulated.