Access Day 1 Highlights
Oct 18th, 2006 by Karen
OurOntario:Yours to Recover - Art Rhyno, and Walter Lewis
OurOntario brings together content and contributed metadata from different sources including newspaper, audio, images, etc. Use Lucene to do this because Lucene is very good at bringing together seperate indexes. Also using Solr to look at facets. All allow Comment on Items (see Great Lakes Images collection for an example). Possibility of using Content-based image retrieval (CBIR), also known as query by image content (QBIC) to find duplicates.
Grid approach to indexing will speed up indexing of records.
Lucene is very well used throughout a variety of systems including the new version of Horizon from Sirsi/Dynix. You can federate Lucene indexes. Offers potential for common index format.
Improving the Catalogue Interface Using Endeca - Tito Sierra
Why Endeca
- Relevance ranking of results
- Browse
- Improved subject access
- Performance/speed
Key Features
- Relevant Ranking
- based on order of modules
- Search and browse in a single interface
- Layered facet refinement
- 11 facets implemented
- spell correction
- "Did you mean…"
- Stemming
- Sort options (publication date, most popular, call number)
Big wins
- Revelance ranking
- Speed/performance
- local managed presentaiton interface
- Persistent parameter based entrypoints (persistent urls)
Features not supported
- Work level aggregation/roll-up
- Customizations/personalization
- Folksonomies/user contributed content
- Reccomender functionality
- Shopping cart functionality
Implementation Challenges
- Deciding which facets to surface as navigation refinements
- Designing user interface
- Optimizing the relevance ranking algorthm
- Optimizing the faceted navigation display
Post Launch Enhancements
- Relevance ranking tweaks
Future Plans
- Aggregated work display
- More browsing options
- Interface improvements and continuted usability testing
- Web Services interfaces
- Search results in RSS/OpenSearch Format
- Catalog Availability Web Service
http://www.lib.ncsu.edu/endeca
Highlights from Lightning Talks
LibraryFind - Jeremy Furmkin (Oregon State University)
Federated searching too slow
Cache searches that people have done in the past
Homegrown federated search tool
Ruby on Rails
Ajax driven User Interface
Two clicks - search and get
Putting Content Online a Practical Guide for Libraries (Chandos Publishing) - Mark Jordan
Wanted to make the book open to the masses
Selected chapters are up online
End to end case study
What do we do with digital content that is too big to fit on the screen? -
MrSID (LizardTech)
Ajax interface that gets information from MisterSid (XML) to figure out what piece of the map to show.
JPEG2000 offers new options
