Library 2.0 September 21st Program Notes

Library 2.0 SIG Business

Clunky Interfaces
Difficult Data Migrations
Software that does some things okay, but nothing well
Can we take our data back from this outdated model?
Representation of Electronic Resources is Generally Unsatisfactory
Sharing data is becoming more important, is the current typical catalog implementation sustainable in this environment?
- Are they accessible to search engines
- Are search results available in anything but poorly formed HTML or Z39.50?
Vendors help us manage complexity - how can the systems we've invested a lot of money in become more...
- More Responsive - Rapid Development Environment
- Integrated with Electronic Resource Data
- With Digital Libaries/Repositories?
- With Metasearch tools, etc.

Interoperability
Support for Web Services like SRU/W, RSS, Opensearch
Management Functions are Tightly Coupled with the User Interface
How do you get a better interface but still..
- Get holdings data into the display?
- Provide a search tool powerful enough for staff power users
the need to index large data sets
marc friendly
character set issues
data normalization - a new interface is likely to show a lot of warts in your data

ILS Replacements see http://liblime.com/
- Solutions for libraries of all sizes?
- Are these ready for primetime?
- Koha A number of public implementations at small libaries
- and the Evergreen Project consortia open source ILS application for the State of Georgia Library System
Overlays
- Add Software to improve search and display functionality
- Generally integrate holdings information after the point of discover - this will need to be formally addressed for these systems to really take off
- Most feature *faceted* browsing

Slow Development Cycle
Long wait for new features
Web 2.0 Technologies enable semi-skilled technicians to rapidly add features and experiment
We need to move to more open, standards-based systems
Open systems generally work with Web 2.0 style technologies Standard 2.0 Demonstration - how long would you need to wait for a vendor something like this?

Examples taken/based upon http://code4lib.org/node/139 - Erick Hatcher Pre-Conference Workshop at Code4lib 2007
Solr Home
Open Source
Java Based Information Retrieval Platform
- Solr is the configuration, results, and management layer
- Lucene is the index tool
- Lucene has been around, Solr is new. Solr makes lucene usable for people in this room
Requirements JDK 1.5 or later, java application server, i.e. tomcat
Comes bundled with Jetty Jave Web App Server to support easy experimentation
Commit|Update using XML, receive search results
- Maybe Open Source is responsive, incremental commits|deletes, the ability to run more multiple indexes at the same time weren't there before.....
Very Fast, does one thing very well SEARCH
Enterprise level performance
- Some ILS vendors are starting to incorporate solr and or lucene into commercial products
A number of User Interfaces have been built, some we've already seen
- Most production services currently for special collections
- Interface Layer is completely separate from the index/data layers
- Easy to have multiple interfaces for different user groups

caption Solr/Lucene

The Update Layer is currently the most Underdeveloped this is what vendors are starting to sell: http://www.exlibrisgroup.com/primo.htm
Data standards make this kind of integration possible

Two config files
- schema.xml
- solrconfig.xml
Ability to Combine Index Fields on the Fly
Wildcard matching in index
Strong support for dates
Highly Customizable - Write you own custom field type, etc.
Multi-valued Fields
Copy Field Values to other Fields (Faceted Display in a number of the systems we've talked about is executed through this type of configuration)
Run multiple indexes on the same solr instance

Select Data You want to Index
- From Catalog
- Special Collections/Digital Projects
- Other Data Sources
Process the Data
- Data needs processed into the Solr XML format
- Simple Key=>Value
- See http://wiki.apache.org/solr/UpdateXmlMessages
Configure the Solr config files http://wiki.apache.org/solr/SchemaXml
Index the data, can be done using the command line and the post.jar file included in the solr distribution
Try out on of the Solr interfaces described at http://wiki.apache.org/solr/IntegratingSolr

Download the software from the solr website http://lucene.apache.org/solr/
Follow the instructions at http://lucene.apache.org/solr/tutorial.html
Potential Pitfall: Make Sure you have a full jdk installed, not just the jre, solr needs a full jdk to run properly