ALA | Design and Implementation of an Open-Source Cross-Database Search Engine

Library & Information Technology Association

Design and Implementation of an Open-Source
Cross-Database Search Engine

Presentation Outline for Scott Ross

The NC LIVE Environment - constraints and liberties
Overview of the NC LIVE "EZ Search" Engine
Decisions Made and Directions Taken
- Z39.50
- Buy it or Build it?
- Sorting, De-duping, and Ranking
- Lexical functions
Content Vendor Issues
- Making a case for getting vendor support
- Vendor native interfaces
- Default behaviors, profiles, and stop-words
- Remote connectivity with and without proxies
Infrastructure Requirements for Implementation
- Hardware
- Software
- Personnel
Localized Customization and Integration
Improvements and Enhancements
- Under Construction
- Planned for Future Development
Conclusion
Questions and Answers

Presentation Outline for Pete MacLaren

Infrastructure Requirements for Implementation

Hardware
- Platform used should reflect the size of the served population and expected load.
- Can be run alongside Apache server or in a separate, standalone server.
Software
- Apache v1.3 (version 2.x has a refresh issue)
- MySQL Database (optional)
- YAZ Toolkit
  1. YAZ compiles on most Unix-systems that with an ANSI-C compiler., the GNU C compiler also works.
  2. For MS users YAZ compiles out-of-box with Microsoft Visual C++. Precompiled binaries of YAZ also available.
  3. Download at http://indexdata.dk/yaz/
- PERL Modules
  1. Event
  2. Net::Z3950 MS users will want to view information at http://perl.z3950.org/support/windows.html Download at http://perl.z3950.org/
  3. Net::Z3950::Record::MARC
  4. Net::Z3950::AsyncZ (If you plan to modify the code)
  5. EZ Search Distribution
- Estimated Time Required to Deploy
  - An experienced Perl implementer should allow for about two weeks to install required software and customization of EZ Search scripts.
  - Additional vendor connectors, that have not previously been written, can be created from scratch in about a week. This is dependent upon factors such as experience walking webpages, and complexity of the vendor’s website, etc.
  - Customizing the HTML output to conform to local environments requires up to a day or more, depending on the complexity of the desired layout and design.