"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 01 April 2008

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


Agenda (Proposed)

  1. Endorsement of Ben's current priorities
  2. Implementation of Profiles in GeoServer
    • Multiple instances of GeoServer or single instance with 'wrapper'
    • What are our options
  3. Schema Validation of Documents
    • Explicit statement of requirements
  4. Timelines
    • GA have no room to change their schedule
    • GSV may have room to extend the deadline a little (the advantage of not having to develop clients)
    • BGS ?
  5. Plans B
    • Query functions - can't see this happening by mid-May
  6. Requirements specification
    • Profiles - need the instance docs to specify them
    • WFS Requests ...
  7. Leave
    • AlistairRitchie on leave in NZ from April 17 until May 6. Bad timing, given the next iteration ends Friday May 9. Do we need someone to fill in as user representative?


  1. Iteration GeoserverIteration2008Apr1 priorities
    • Confirmed by group.
  2. Implementation of Profiles in GeoServer
    • To deliver the same FeatureType using different profiles requires different services.
    • Multiple instances of GeoServer or single instance with a lightweight 'wrapper'
    • BGS used multiple instances of GeoServer to implement Testbed 2 and may well do the same for Testbed 3.
    • GSV will investigate URL redirection and use of simple servlets to allow multiple 'services' to be implemented with a single GeoServer instance. (More on this to come.) GA (possibly BGS) will consider the same approach. Neither GA or GSV have web servers that could handle several GeoServer instances. This is a site-specific implementation decision, not a work-around for a perceived GeoServer bug.
    • Action AlistairRitchie to distribute emails discussing this issue to TimDuffy and DalePercival. - Completed
    • Action GSV to investigate and undertake the development of a Java wrapper servlet and make available to GA, BGS (if required). Subject to approval of management.
  3. Schema Validation of Documents
    • Confirmed that responses to requests against Testbed 3 services must be schema valid.
    • Need to confirm: can we accept just valid GeoSciML (and other community schema) fragments in the document/response or does the whole response have to validate. I.e. the GeoSciML and the WFS wrapper. wfs:FeatureCollection elements etc.
  4. Timelines
    • GA have no room to change their schedule, but will consider using GeoServer builds released after May 9 for Testbed 3 if they are suitable and stable. They will be working on their own bespoke WFS to fill any Testbed capability gaps left by GeoServer at this time.
    • GSV have room to extend the deadline into June but are not sure at present exactly when. They will investigate the use of the Geological Survey of Canada's multi-tier WFS solution as a back-up plan.
    • BGS are in a similar position to GSV.
    • Stable Testbed 3 services must be delivered by June 30, 2008.
  5. Plans B
    • As stated above. GA: their own WFS; GSV and BGS to investigate the GSC multi-tier approach.
  6. Requirements specification
    • Profiles.
      • The Enterprise Architect Testbed 3 Profiles project (insert link) will be used as the basis for establishing content.
      • Annotated instance documents are to be compiled for each Use Case profile.
      • TimDuffy to allocate various Testbed participants to the task. (GSV to provide a Use Case 4 - Geomodeller profile by August 16.)
    • WFS Requests. Profiles for these required for Use Case 3n. (Query profile.)
  7. Leave
    • AlistairRitchie (Bruce - Victoria) on leave in NZ from April 17 until May 6. Bad timing, given the next iteration ends Friday May 9.
    • DalePercival (Bruce - Australia) will take the rule during this period (thanks! -- AlistairRitchie).
  8. Other issues arising
    • TimDuffy wants to know what the status of the Shapefile datastore is.
    • Need to know how formal notification of the release of a GeoServer build is made.
    • Clarify the use of GeoTime vs. ControlledConcept for GeologicEvent\eventAge values in Testbed 3.
      • Testbed 3 group assume that the descision to use a URN pointing to a ControlledConcept vocabulary as the CGI_Value (term) value(s) still stands.
      • Our understanding is that OliverRaymond (and possibly SimonCox) expects GeoTime to be used for Testbed 3.
      • Reasonable expectation, but the Testbed Group (including the designers of the BRGM vocabulary service) do not have the resources to actually accommodate this new requirement.
      • Action Testbed group to reiterate the use of URNs and ControlledConcept for GeologicEvent\eventAge values and recommend a review of this for GeoSciML vN.N.
      • Action DalePercival to discuss this issue with OliverRaymond.
Topic revision: r8 - 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).