[Devforum] November 18th Meeting Minutes

[Karen Reardon]

Meeting convened and minutes taken by George Ouellette:

Developers Forum 11/18

Present: Jeff Barnett, Matthew Beacom, Art Belanger, Meg Bellinger, 
Richard Boursey, Roy Lechich, Scott Matheson, Teresa Miguel, Audrey Novak, 
George Ouellette, Kim Parker, Kalee Sprague, Joan Swanekamp, Jennifer 
Weintraub, Stephen Yearl

Audrey updated the group on her groups work with OAI, Metalib and Orbis.
Some key points:
The process of harvesting sub-sets of Orbis records has begun. The group 
is focusing on finding aids first but also plans to include E-books, 
videos, and new books.
A challenge that they will meet is the process of identifying the records 
that will belong to each subset.
To harvest a sub-set of Orbis records a local table must be created 
containing the records to process, dynamic harvesting of sub-sets isn't 
possible yet.
Data doesn't need to be harvested to work with OAI, formed records (i.e. 
XML) will also work if they are contained in an "OAI Wrapper"
The goal of the harvesting is to bring in records that are not Z39.50 
accessible.

Jeff the spoke about an AOI post-conference he attended during the DLF 
conference.
Some key points:
Some of the challenges were brought up for the sharing of metadata across 
institutions were lack of consistency between institutions, lack of 
contextual information, lack of collection level information.
To address the consistency issue it was suggested that sites use a 
standard set of core data along with a site specific extended schema to 
fulfill local needs. It was also suggested that sites map their data early 
in the request process rather then transforming the data at the end. Also, 
the XML source should be validated before the harvesting process begins.
There are a series of different tools available for serving and harvesting 
depending on the repository and site.
Metadata migratory tools are available to aid in the migration of data.
MOD-OAI allows the serving of OAI data on an apache server. It is being 
used by one institution but we don't know if it is a release or beta.
Google doesn't harvest metadata, it harvests the URL from the metadata and 
crawls the referring page.

Art Blanger updated the group on some tests he has conducted with the 
Greenstone export of OAI data. He reported that it works on the test 
instance the is bundled with Greenstone but he hasn't tried it fully with 
their production data.

Scott Matheson reported that BPress has the ability to be OAI harvested 
and while it only has about a half-dozen records that he is willing to use 
it in test cases.