Science Environment for Ecological Knowledge
Ecoinformatics site parent site of Partnership for Biodiversity Informatics site parent site of SEEK - Home
Science Environment for Ecological Knowledge









 

 

 



Kepler Conference Call Notes 19 Aug 2004

Participants

Jones, Tao, Ludaescher, Berkley, Spears, Pereira, Altintas

Discussion

1. Ecological Niche Modeling pipeline

    • Ricardo gave summary of Open modeler, suggested it be incorporated into Kepler
      • framework for modeling algorithms, geospatial sampling, etc
      • better base for ENM models than current GARP code in Kepler, more OO, has Java bindings, better pipelines, new algoritms would be available automatically
      • current algorithms in OM are GARP, CSM (Climate Space Model based on principal components analysis)
      • Ricardo would like to use Kepler as the UI for OM, and is willing to integrate into Kepler -- this may solve our issues for both pre- and post- processing of the geospatial data
      • need to expose components as standalone actors to whatever extent possible
      • Ricardo will try to implement by Oct 15 with help from Dan (Chad until Dan returns from vacation)
    • GIS actors status
      • GRASS web service completed but needs to be deployed
    • image display actors for GIS data -- outstanding needs?
      • Open modeler provides capability to get both actual image results plus low-res images for display, but Ricardo says only really useful for the ENM stuff -- probably better to use GDAL

2. EcoGrid and data access in Kepler

    • Rod gave overview of QueryBuilder and planned integration into Kepler
      • Might be hard to have the dialog appear when the object is dropped on the canvas, so might need to come up with another model. But need to investigate initialize() further
      • Bertram suggested ability to use QB as a 'SQL Filter' in the middle of a pipeline -- we agreed it was a good idea and we should create a task for this, but it would be secondary in priority to getting the EcoGrid side connected
      • Rod showed the QB html page, and asked for feedback -- he'll be asking the SEEK users for feedback too
      • Bertram brought up that a design tool might also be nice (DBDesigner4)
    • Metadata query facility changes needed
    • Data query implementation status
      • Jing says it is close to complete for the hsql embedded db
      • Confident we can get the execution engine doen by the Oct 15 deadline
    • Mammal data access -- and metadata -- DIGIR results status
      • data is probably available from MANIS DiGIR network, as well as GBIF
      • need to update the DiGIR ecogrid implementation to use the registry (possibly Bing?)
      • Matt will organize a conference call fro EcoGrid to resolve these issues
    • Environmental data layer access -- and metadata
      • who has these (IPCC?) and how do we get them into the EcoGrid with metadata?
      • Ricardo says that they are getting away from IPCC data -- we need to find out
        • Matt will contact Deana about environmental data

3. Ontology-based browsing for actors and data

    • Chad implemented search interface for the actors, pluggable to accomodate new ontology search interface
    • Chad will contact Shawn about further development on this, in particular
      • need mechanism for semantic registration for MoML
      • need access to sparrow API via Java
      • Chad still working on trimming the result tree to exclude all non-matched items and categories
        • Wants to implement the double-window view to compare effectiveness
    • Chad estimates that Oct 15 delivery date is possible (he's going on vacation for a week though)

4. Saving kepler products to EcoGrid

    • Prioritization of 'put' interface work
    • Matt: need ability to archive data products with their structural and provenance metadata
      • Bertram: do we have one systematic way to store metadata about execution (execution trace) that includes parameter settings, versions of services, etc.
        • Assumes notion of global ID's for services and data objects
    • Matt: also need ability to save workflows themselves in Grid environment, for sharing and collaboration
      • Then users can search and find components others create, including web services and other services that are Grid-available
    • Ilkay: 'Blue-box' recorder (parallel to director) that decides what needs to be recorded and archived
    • Bertram: maybe short-term is to get actor to write back to EcoGrid nodes and can be incorporated into workflow
      • Ilkay says the web service harvester may be incorporated by Oct 15 to allow finding remote services -- possibly with help from Bing and/or Efrat
      • Ilkay also says we should trim the tree for the release to only include stable stuff

5. Kepler hackathon -- should we have another

    • Ilkay says it was useful. Would like another. But we had a hard time finding an available time. The week of Oct 11-15 might be possible, but it is right between 4 other weeks of meetings, and is probably too close to our intended Oct 15 release date. We also have the SEEK meeting as a possible planning meeting in the first week of November. Maybe we could ask more Kepler people to come to at least a portion of that?
    • Matt and Bertram will follow through on whether the SEEK meeting is the right place for the next Kepler planning meeting



Go to top   Edit this page   More info...   Attach file...
This page last changed on 19-Aug-2004 12:50:37 PDT by NCEAS.jones.