NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.
Hi Jon, All,
I don't know if considering profiles of extensions is just complicating the
issue though... I think the reality of the core + extensions approach means that we will *have* to agree on supported combinations of extensions to achieve interoperability. i.e. agree to provide a WCS service which consists of: Core + ExtensionA + ExtensionD and ExtensionF So in that sense I think we need community profiles of WCS, but if it becomes necessary to profile an individual extension (which is what I think you are implying?), then I think the extension probably needs breaking down into simpler more modular extensions. i.e. I don't think we should be trying to define how to do interpolation or subsampling within the WCS_NetCDF extension*. Interpolation should be a separate extension. i.e. it would be good to have this model: Core + WCS_NetCDF + Interpolation Rather than this one: Core + WCS_InterpolationProfileofNetCDF If the WCSNetCDF extension is in itself too complicated to implement (fully) then that probably needs addressing at this stage before it becomes an official OGC document. Cheers, Dom *we're not, this was just an example! On Wednesday 01 October 2008 13:54:07 Jon Blower wrote:
Brief replies to John's notes:because the data volumes can be so large, we have found that visualization users want subsampling.Good point, I mustn't forget that WCS can also supply visualization applications.I assume by curvilinear, you mean ones with 2D lat/lon, such as swath data, not the simpler case of projection coordinates.Yes, that's what I meant, sorry for the confusion.Can anyone who knows WCS 1.2 summarize 1) whats in the core, and 2) what is allowed in the extensions?I can't help here, but from previous posts it seems that the core defines no output formats at all and therefore needs to be combined with at least one extension to function at all. (To my mind this is a misuse of the word "extension" but never mind.) So it looks like my two categories of user don't map neatly to the WCS core+extension model. It might still, however, be valuable to consider "profiles" of WCS-NetCDF for different categories of user, for practical reasons of focussing effort and "doing the easy stuff first". I don't know if considering profiles of extensions is just complicating the issue though... Jon
galeon archives: