|
- Present: Gauch, Stewart, Gales, Beach, Jones
- Discussion of the status of the current concept IR demo
- Challenges getting PHP SOAP libraries to work
- Discussion about which languages we should be using in the future, conclusion that whichever language is best for a particular subprojecct will be fine. Susan likes C++ for speed, Python for ease of use and available libraries.
- Some discussion of RDF and looking at that for representing concepts.
- Connected in Matt Jones on the call to apprise him of current status and short term prototype development plans.
- Discussion on whether concepts in data sets need to be indexed centrally in a database, Matt thought it would eventually not be necesary, Susan not so sure. Not an issue at the moment. Susan: we will need a very quick way of determining which data sets have which names in them. A pre-computed concept index would provide rapid query responses, the possibility of querying each (metacat) data set for each name for each query seemed onerous to the KU crowd.
- In the end, the current plans for the elaboration of the IR demo seemed fine to Matt, Susan and all. KU will purchase and install a small Linux server to host our prototype services and two APIs (1-Resolve-Concepts, 2-Compare-Concepts.)
- Stewart summarized the current assumptions, definitions and status of the taxon IR demo here: DefinitionsAssumptionsAndGeneralArchitecture and PlanOfAttack, and also started editing and hardening the use cases from the January 30 2003 meeting here: UseCases
|
