Changes between Version 4 and Version 5 of WorkingGrid/DataProviderSetup/PML/NOCSJun07

04/06/07 13:02:14 (12 years ago)



  • WorkingGrid/DataProviderSetup/PML/NOCSJun07

    v4 v5  
    44 * Determine precise staffing situation before Board Mtg, how much of Mike/Mark/Me we have left, to do what, by when, and PML case for extension. 
    55 * We need to get Browse access to our Exist DB working.  
     6 * We need DataGranules to be properly included in our MOLES, so that Bryan's Browse can collect them in Shopping Cart to pass through to Visualise. 
     7 * We need to reduce number of discoverable records to be more useful in Discovery, and avoid flooding results. 
     8 * Possibly use multiple bounding boxes (and multiple time ranges) to represent our multi-region datasets. 
    711== Financial and staffing == 
    1317== Discovery and browse == 
    1518 * Went through current GUI's, issues, B metadata connection problems e.g. firewall. 
    1720== Vocabulary services == 
    1921 * Demo at but no longer publicly linked. 
    2123== CSML 2 == 
    2324 * PML work (Mark) on CSML plugin for Raw image data is working and is first example of a new format being plugged into CSML. 
    2425 * Showed working demo of converting model output NetCDF file into a CSML record XML using a scanner, then a small Python program to parse the CSML doc to extract entire gridseries and subsets in space/time, profile-series. The output is a further CSML file to describe the subsets, together with the output NetCDF files, viewed for now using 'ncview'. 
    2829== Visualisation == 
    3030 * Steve explained OWS system developed at BADC for viewing CSML data via OGC methods: OwsFramework, OwsNotes. 
    3131 * Example OGC WMS client showing IPCC data served via CSML: 
     32 * Bryan says they will tackle the PML+Model vis use case first, then the Model+BODC CTD profile use case. There will be a 'Shopping Cart' facility from Browse to pass through to Vis. 
     33 * For that PML will need DataGranules to be working in our MOLES. 
     35== Security == 
     37== Metadata population == 
     39 * PML 'blocker' is linking MOLES to CSML granules. 
     40 * Reduce number of PML DIF records, e.g. All MODIS data for each region? Or for a whole Europe region (ie one record per sensor)? 
     41 * Multiple bounding boxes in MOLES, and possibly in DIF - available but not used yet. We could use this to reduce DIF records, then the Browse would need to pre-populate the Granule browse with only those within the searched region of interest (Bryan says that is possible). 
     42 * I had my usual argument with Bryan on intelligent ordering of discovery records so it can handle any granularity. He said he was intending to assess the size of requested and dataset regions, and prioritise smaller matching regions above global datasets.