Minutes of Telecon 20 June 2011

Attendees

  • Russ Dengel (SSEC)
  • Stefan Siemen (ECMWF)
  • Ben Domenico (NCAR)
  • Michael Weis (IBL)
  • Matthias Mueller (Univ Dresden)
  • Frederic Guillaud & Marie-Francoise Voidrot (Meteo-France)

Proposed Agenda :

  1. Egows I.E. Feedback
  2. MetOcean Best Practices improvements
  3. Others
  4. Next Telcon
1) Egows I.E. Feedback :

On the 9th of June participants from the European Group on Meteorogical Workstations have made some informal tries of interoperability between several servers and clients
  • Tested servers: IBL, Meteo-France, ECMWF, Univ Dresden, UCAR
  • Tested clients :
    • Clients developped by Met Ocean data experts : Ninjo (DWD) , Visual Weather (IBL), KNMI, Metview (ECMWF), Synergie (Meteo-France) ,
    • Open source general purpose clients : gvSIG, Gaia
A synthesis is under work because we still expect more screenshots.

The day began with free tries to test the means available. Then we tried to see hao we could validate the responses from the servers by combining the same type of data for the date date from different servers.

This could be done with GFS data served be the Ucar server and the IBL server , Radar data served by IBL server and Meteo-France server for instance.

We found again the issues already identified of styles but also thresholds on the radar images.

We then tried to implement the use case defined several monthes ago by forecaters based on the combination of Mean Sea Level Pressure fields out of different models.

This field was available out of the GFS mode on the UCAR server and the IBL server and out of the ECMWF model on the ECMWF server.

The capabilities are long and the name not homogeneous : so beside the issue of standardising the names of the layers already identified, appeared the need of having the possibility on the clients to order or look up for names.

We have met a problem with the style of the UCAR due to the range of values of the MSLP that didn't fit the default one. Ben provided complementary information on a server option COLORSCALERANGE described there : http://www.resc.rdg.ac.uk/trac/ncWMS/wiki/WmsExtensions

We then tested the Univ Dresden server that provides grid coverages with monthly projections of temperature, precipitation and cloudiness from january 2006 till december 2099 for different SRES scenarios.in order to involve climate tests into our works.

Matthias had provided these guidances to the tests :

"We use the start/stop/period syntax from WMS 1.3.0 in the capabilities, provide a default legend for each of the variables and a simple getFeatureInfo xml. What we would like to test is the compatibility with existing clients according to:

1) time - can current clients interpret the start/stop/period syntax for monthly data?
2) legends - are they readable by the clients and self-explanatory to ad-hoc users?
3) getFeatureInfo - can the featureInfo XML be interpreted and displayed by the participating clients?
4) delivery speed - is it good/bad/sufficient for end-user and expert clients?

What we expect from the testing (besides a collection of successes and fails) is a list of issues that currently hinder an ad-hoc use and a first-glance interpretation of the maps (concerning the time semantics, the legends, the FeatureInfo and the delivery speed)."

We found some troubles with the styles names that included "/" (sends back an error. (internal server)).

=> / : ! Are to be avoid in layer or style names or identifier
With no style the map is displayed


Time Start / stop period was ok with several clients. If not it was considered as a problem of maturity of the clients because the syntax was correct.

The Legends were readable
Response time : Time for rendering seems under 3seconds so Good then the network delays were not linked to the server.

gvSIG handles the time which is rare on general purpose clients. It is a 1.1.1 implementation.

Last, we concluded that we should define a checklist of tests . "cite.opengeospatial.org" already provides a good base of automatic testings to validate a server.

2) MetOcean Best Practices improvements

Chris, Stefan and Marie-Francoise propose to make a first draft of WMS Met Ocena Best practices that will follow exactly the map of content of the WMS1.3 spec. The first aim will be to make a BP doc for WMS1.3 because it is there and recommended by INSPIRE. Then we'll see if it can be extended to other WMS releases. Following the organisation of the spec document should easy the joined lecture by developpers and the contributions by all. . The BP doc should provide examples based on Met Ocean data.

We encourage all volunteerings to send contributions and examples of implementations with a reference to the relevent paragraph.

The next teleconf should discuss a proposal for handling TIME that should be provided by Marie-Francoise and the following one should discuss the layer names standardisation. Works are going on on this subject into the INSPIRE Met experts group. A link to their work will be available a this time.

3) Others

Several members of the Met Ocean DWG contribute to the outreach of the works. (Russ in NWS, Dave Santek at a Eumetsat conf soon, Marie-Francoise probably at the AMS). A share of the presentations could easy this work and make it more consistent.

4) Next telco

Next week same time ont the Tme handling Best Practices

-- MarieFrancoiseVoidrotMartinez - 20 Jun 2011
Topic attachments
I Attachment Action Size Date Who Comment
Synthese.pptppt Synthese.ppt manage 7410.5 K 20 Jun 2011 - 14:43 MarieFrancoiseVoidrotMartinez Synthesis of EGOWS Interop testing
Topic revision: r2 - 22 Jun 2011, MarieFrancoiseVoidrotMartinez
 

This site is powered by FoswikiThe information you supply is used for OGC purposes only. We will never pass your contact details to any third party without your prior consent.
If you enter content here you are agreeing to the OGC privacy policy.

Copyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding OGC Public Wiki? Send feedback