NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.
Hi Peter,
wouldn't it be nice to store NetCDF and get out GeoTIFF slices at some time?
Yes! IMHO think this will do more for "GIS interoperability" than working out how to have NetCDF as an *output* format. (Again, I repeat my worry about the "core plus extension" model...) Regarding your other points, yes, with enough effort we could figure out how to force CF-NetCDF through WCS without loss of information (although WCS will be years behind the curve of scientific innovation for the foreseeable future unless we get Roy's suggested $700bn bailout). My argument remains that this isn't what we really need. We need to present metocean data in a form that others can understand. Having another way to download CF-NetCDF doesn't help those who can't understand CF-NetCDF. Cheers, Jon On Mon, Sep 29, 2008 at 1:55 PM, Peter Baumann <p.baumann@xxxxxxxxxxxxxxxxxxxx> wrote:
Hi Jon, thanks for explaining to the newbie in metocean data (saw some of the points in earlier postings). To me it looks like there are some issues which can be treated individually: - data not on regular grids. This is an issue the WCS group is aware of, but up to now we haven't found anybody of expertise who is willing to contribute a spec. I now have a PhD student sitting on this. While he has a sound geodesy / remote sensing background he will need a little time to get up to speed. This work I expect to close this gap. - non-standard CRS. One option WCS offers is to use just the Image CRS, i.e.: the integer array coordinates if you will. If this CRS is the only one the WCS offers then no detailed handling is necessary - DescribeCoverage offers some identifier, GetCoverage uses it. BTW, you can effectively prohibit interpolation this way (a requirement I have been fighting for). - translation of subsetting. Again, a coverge description contains a boundary box. If your WCS provides that then the GetCoverage request can contain this box or an extent inside thereof. 2) The ocean community lacks a means of describing coverages in an interoperable way (which is one reason why Stefano's document will be important). Hence I think that WCS GetCapabilities and DescribeCoverage could be useful, even if the coverage is ultimately accessed by other means. Not sure that the outcome of this construct is really interoperable. As earlier the connection into the GIS user community was mentioned, wouldn't it be nice to store NetCDF and get out GeoTIFF slices at some time? Norman (or anyone else): how does this relate to the JPEG2000/JPIP situation? Does a JPIP stream point to an entire coverage or a coverage subset? JPIP has its problems when it comes to general subsetting. I personally can imagine that either they support that by all means (which will sometimes mean: bad performance) or they refuse and return an exception on parameter constellations they don't like. But that's my personal opinion. Anyway, good discussion; I feel this helps to increase udnerstanding (at least from my side). -Peter
galeon archives: