[galeon] TDS WCS 1.0+ implementation and questions

NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.

Hi all,

[Sorry for the cross post.]

I have implemented two new WCS versions in the TDS. The first one (version="1.0.0.1") makes some improvements on our current 1.0.0 server. Mainly it allows for min/max selections on time and vertical axes (no resolutions on min/max and no lists) as opposed to the single point on the time and vertical axes allowed previously. There are also some error handling improvements. I also dropped GeoTIFF support for now.

The second one (version="1.0.0.11") implements some of the WCS 1.0+ goals. Mainly allowing for multiple range fields in a single coverage. One issue here is that the coverage names are kind of ugly, e.g., "time isobaric y x". The description is a bit better, e.g., "All the parameters on the 'time isobaric y x' coordinate system: Temperature, ..." but still kind of ugly.

My summary of the changes from 1.0, issues, restrictions, and such is at

http://www.unidata.ucar.edu/staff/edavis/notes/WCS/TDS-WCS-1.0+_Impl.html

I'm afraid we don't have either of these new versions up on our public server (hopefully we will be adding a public development server next week). But to provide some examples of the changes listed at the above URL, I do have a few test requests and responses at

http://www.unidata.ucar.edu/staff/edavis/notes/WCS/TDS-WCS_Testing.html



My main question has to do with CRS. I'm not sure of the best way to deal with our "3-D with weird vertical axis" CRS. I'm leaning towards providing URLs to a GML encoded CRS. (A bit more detail at the summary page listed above.) Does that seem like the best route? Are there other GML CRS profiles I should look at?

My other question has to do with the DescribeCoverage rangeSet. I went with multiple 1.1-ish Field elements instead of Andrew's suggested extension of 1.0 to allow for multiple axisDescription elements. (Again, more detail in the summary listed above and an example at the test URL above.) But that decision was mainly just to get something working and to include data type and units. Any thoughts on directions to go with rangeSet?

I think that is it for now. Comments welcome.

Ethan

--
Ethan R. Davis                                Telephone: (303) 497-8155
Software Engineer                             Fax:       (303) 497-8690
UCAR Unidata Program Center                   E-mail:    edavis@xxxxxxxx
P.O. Box 3000
Boulder, CO  80307-3000                       http://www.unidata.ucar.edu/
---------------------------------------------------------------------------




  • 2007 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the galeon archives: