Notes from the presentations at the Catalog Form & Function Interest Group, January 16, 2010 from 10:30 am to 12:00 pm.

Recent Trends in Catalog Architecture

Univ of Chicago Library – 6 million bibs

Uses a program called “lens”, which is compatible with Aquabrowser I think. Lens allows users to perform a full db search as well as faceted. Part of its UI lists the full counts for each part of the catalog itself. This may be a Horizon system-based program which uses Linux I think since the speaker kept mentioning “cron jobs”.

Includes NAF & SAF integration, though the NAF isn’t quite ready for prime-time. Also experimenting with Oxford Dictionary of National Biographies. Uses fulltext Google searching along with Halthi.

Speaker compared it to Stone Soup—“bare bones of data from ILS but with extra stuff mixed in.”

They cancelled TOC from BNA (no reason or when), so they are harvesting it from Syndetics now. Incorporates SFX & EAD.

Tried harvesting Mbooks to Horizon, basically very brief headings (110 $aUnited States), but had some difficulty. Users of ProQuest/EBSCO. Their records are not natively in Unicode so they have to run special conversion programs to get the records to load into lens.

http://lens.lib.uchicago.edu

Discovery Layers

Michigan State University - Joshua Barton

They stated that their presentation was just an idea they were floating around, still looking into ways of collating / accessing their data with more success.

Interested in a next-gen catalog interface that could be a one-stop-shop for their catalog, digital collection, and articles in their db.

They use III’s Encore system. Apparently they need to massage the data for every bib they load, which they find is tedious. There is no alert when authorities are updated. The Name & Subject authorities are not synched between two databases (digital & catalog) so overlap / duplicates exist in their auth db.

They would prefer a modular system that makes it easier to add / remove certain aspects of the program.

However, if they change their existing configuration for what they think they want (modular layers), they will lose:
• Metadata customization
• Controlled vocabulary
• Normalization capabilities

They’re not even sure if what they want is feasible or will work. Talked a little about the differences between Aggregated vs Federated searching.

Automated Metadata - Repurposing Using eXtensible Catalog Software

Univ of Rochester - Jennifer Bowen

Working with a grant from the Andrew W Mellon foundation, Rochester is developing an open-source toolset for two reasons:
• Resource discovery
• Metadata management

This is considered a work-in-progress, though it appears to work rather well for small batches (~50,000) of files.

The UI utilizes a faceted, FRBRized approach (search) that users can view via web interface. The service offers various tools to assist in automated metadata processing from several different (XML) sources. It uses a standard protocol called OAI-PMH that has the capability of other users in mind.

Its frontend is: Drupal CMS. It will convert MARC21 to MARCXML.

http://www.extensiblecatalog.org/

Summon

Univ of Calgary - Aaron Wood

UCal is a beta partner with SerialSolutions for their Summon software as of 2009.

Summon relies on relevance ranking that determines matches based on an internal weighting system. It uses a facet-oriented search, setting aside a total of 56 fields (includes fixed fields) for MARC mapping and identification to users. The mapping module was actually codeveloped by Aaron and another librarian at UCal.

Aaron did not like Dublin Core and said it dumbs down the MARC format. Its OAI Harvester puts everything (controlled vocabularies, etc) in generic descriptor fields, thus removing any significance they might have had previously.