"Seegrid will be due for a migration to confluence on the 1st of August. Any update on or after the 1st of August will NOT be migrated"

Geoserver Meeting 02 June 2008

A Skype conference call between GeoSciML Testbed3 users to discuss GeoServer development.

Attending

Agenda

  1. Consistency of profile instance documents
    • Is GeologicUnitPart mandatory for UC2A?
  2. Discussion of known issues with GeoServer
    • GetCapabilities; DescribeFeatureType; maxFeature
    • Population of @codespace attribute values
  3. Recursion in responses (eg. GeologicUnitPart)
    • Serialisation in-line vs. href.
  4. Test data for Use Case 2A
    • Sample datasets to match instance documents
  5. Use Case 3
    • Structure of of WFS filter queries
    • Implementation plan.
  6. Plans B
    • How to handle errors in 2.
  7. Enny other business

Minutes

  1. Total: 93min 30s
  2. Consistency of profile instance documents
    • Instance documents for TB3 profiles need to be consistent to be of any value to BenCaradocDavies - needs to know what is to be returned
    • Clearly show all elements expected for a full response for the profile, and which are mandatory
    • Structure must be consistent between documents for the some use case posted by different organisations
    • For the instance documents if an organisation can't provide data for an element (mandatory or optional) either include a comment to show the element is missing or use a NULL value
    • Is the CGI-GeoSciML_Testbed3_Profiles.eap EA project the basis from which profile instance documents should be demonstrated, and if so, is this what the instance docs will be tested against?
    • Action AlistairRitchie to provide comparison of documents for UC2A
    • Action TimDuffy to get JohnLaxton to review instance documents for content and consistency as they are posted
  3. Discussion of known issues with GeoServer
    • GetCapabilities (GEOS-1929), DescribeFeatureType (GEOS-1928) and maxFeatures (GEOS-1930) don't work
    • @codespace attribute values are incorrectly populated, even when the GeoServer mapping file is correctly populated
    • Decision Will not try to have this resolved in time for Testbed 3. Development will continue on UC2 and UC3. Participating agencies will investigate ways of trapping for these errors. Possible solutions:
      • GetCapabilities (URL...), DescribeFeatureType: redirect requests to a static xml document (thanks Eric)
      • @codespace: use XSLT to 'post-process' GeoServer output
  4. Recursion in responses (eg. GeologicUnitPart)
    • Nesting of hierarchical features (eg. GeologicUnitPart) in-line in documents is extremely difficult to do in GeoServer - very difficult to write the SQL for the mapping file.
    • Why not use href? Consistent with typical web applications.
      • Because no clients can resolve these references. (But if we can't serialise them in the first place ...)
  5. Test data for Use Case 2A
    • Ben needs test data for each use case. These data have to match the instance documents provide for each use case. Can then use the instance docs to confirm that he has delivered the sample data correctly.
    • Action AlistairRitchie to organise a Skype conference call involving the User Group and BenCaradocDavies to discussthis issue (and others).
      • Tentative time for this call: Wednesday June 04 at 1700 Aus EST, 1500 Aus WST, 0800 BST
  6. Use Case 3
    • Do we work on UC3 or solve all known issues (pt 3) first?
    • Decision Work on UC3. Investigate solutions prosed in pt 3 as a temporary fix for the known issues.
    • Need to finalise syntax of WFS filter query functions. Esp. deal with questions raise by EricBoisvert on TestBed3UseCase3AProfile.
    • Action GA to provide their expected function calls for each of UC3A-D (or part thereof) and compare with Eric's proposal.
  7. Other business
    • Reviewed JeanJacquesSerrano email of June 02 '08.
    • Reviewed debate on ControlledConcept vs GeoTime for AgeInterval values again.
    • Testbed members in this meeting recommend that we should continue with current proposal to populate term values with urns pointing at a controlled concept until the Model Design Group provide clear resolution/definition of the model to be used:
      • AgeIntervals as ControlledConcept with a GeoTime.TimeOrdinalEra prototype; or
      • AgeIntervals as GeoTime.TimeOrdinalEras
      • Big problem(?): Do we have a data type that handles an eventAge as an AgeDate (numeric value) or an AgeInterval defined by either a ControlledConcept or GeoTime.TimeOrdinalEra?
    • Group recommends final decision on model to be made after Testbed 3 demo by Model Design Group (at Uppsala?). (We demonstrate v2.0-RC1 and then fix and release v2.0.)
    • Action TimDuffy to confirm decision and communicate it promptly to the Testbed Task Group and Model Design Group
    • Action TimDuffy to communicate decision to OliverRaymond
Topic revision: r3 - 15 Oct 2010, UnknownUser
 

Current license: All material on this collaboration platform is licensed under a Creative Commons Attribution 3.0 Australia Licence (CC BY 3.0).