2008-05-21: Subcommittee on Interoperability Telecon Minutes
Back to < Subcommittee on Interoperability
Subcommittee on Interoperability[edit | edit source]
Ad Hoc Advisory Committee -- Air Quality Data Summit
May 21, 2008, 2:00 – 3:00pm EDT
Conference Room C351K
Conference Call Meeting Notes[edit | edit source]
COMMITTEE MEMBERS, Attendees 5/07/08 bolded:[edit | edit source]
Tim Dye / Steve Ludwig
NEXT CONFERENCE CALL:[edit | edit source]
Wednesday, June 4, 2008 from 2:00 – 3:00pm EDT. Conference calls are scheduled on a bi-weekly basis.
MAJOR DECISIONS / CONCURRENCES:[edit | edit source]
- REST binding (there is no normative SOAP binding for WCS anyway)
- Messaging system for WCS is key value pairs
- Preferred payload types are netCDF and CSV [LS: does XML schema have merit?]
- AQS and Airnow will pilot a WCS interface.
ACTION ITEMS:[edit | edit source]
- Need to start a list/use one in existence (NARSTO/CF Conventions) of standard parameter names
[LS: * need to settle on an initial list of methods as well]
- Create introductory/tutorial material on WCS (suggested 10 slide ppt or screencast with Intro, block diagrams and discussion)
[LS: I think the intro/tutorial would be on the "big picture" not just WCS right?]
Summary of discussions[edit | edit source]
Call started with an over view of WCS from Rudy Husar. He started with two basic questions: (1) What is the nature of the query for data? (2) How do you get the data back?
In WCS coverage = parameter and there is a three part sequence between client and server to request data. First client asks for GetCoverage for a particular dataset and the server returns all parameters. Then the user picks a parameter and asks the server to DescribeCoverage. Finally, the client asks for the data with the GetData call. Sig would like at some point to also discuss how you implement. Rudy briefly described the data wrapping process for three classes: (1) SQL (2) fixed structured files and (3) netCDF.
WCS is the lowest common denominator for interoperability by allowing one parameter to be called at a time. To clarify the layer question from two weeks ago, one layer is one parameter from one dataset.
The group had a discussion about converging on standard names for the parameters, but decided that there would always be a need for a mediator to link synonyms to a standard naming convention.