Eco Grid Conference Call Notes 17 Sep 2004

This is version 15. It is not the current version, and thus it cannot be edited.
[Back to current version]   [Restore this version]

Participants

Agenda

This is a continuation of the previous agenda.

1. Certificates and authentication
• see CertificateAuthorityDesign for background
2. EcoGrid query semantics
• handling EML version changes (2.0.0 vs. 2.0.1)
• translating EML and Darwin Core metadata queries
3. Update on data query/subset functionality (Tao and Spears)
4. Update on put interface implementation (Zhu)
• Review the 'put' interface via its WSDL
• Review the 'Authentication' interface via its WSDL

Notes

1. Skipping certificates agenda item because brunt and murillo absent -- will reschedule
2. Ecogrid query semantics
• need ontology mapping between standards
• can Sparrow provide the necessary reasoning services?
• probably need to translate metadata instances as well as classes, so may need to populate the reasoner with al of the instance data (e.g., load EML into the reasoner)
• can existing query rewriting systems work? maybe...but probably would need to be customized or extended
• need to discuss with Bertram and Shawn, especially with regards to Sparrow -- this is probably part of SMS layer, but how to deal with instance data (e.g., EML values)
• probably start with just crosswalks, then add instance data later
• implement this as an 'ontology server'
• e.g., takes an EML query in, gives a Darwin Core query back
• Shawn: says we should map to an integrated schema to avoin n^n mappings
• e.g., in ant example paper they map to the ontology
3. Update on data query/subset functionality (Tao and Spears)
• Rod gave an overview of the use of EML Data Source and cached queries
• Matt: can we use the existing Kepler-side infrastructure to implement EcoGrid-side data query by loading the EML-described sources into a relational db
• Jing: still need the DarwinCore Data source
4. Update on put interface implementation (Zhu)
• Bing gave overview of put implementation
• Noted that Metacat identifies metadata and data separately while SRB uses same id for both

Action items

1. Matt will contact Bertram and Shawn about query translation needs, involve other EcoGrid folks
• reread Shawn's 'ant' paper for ideas
2. Rod will describe (in a bug) and implement a DarwinCore data source
3. Jing will describe and implement a new multi-node EcoGrid search for Kepler (with resultset integration)