NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.
Hi Jon, you are absolutely correct in mentioning ISO 19123 which is our "bible", even though we feel that it would deserve some upgrade. Raster coverages have been our starting point - and still keep us amazingly busy - but in the end our goal is to support all coverage types. ...and then you are right in that the terms coverage and feature are not really clearly distinct in their meaning. Up to now we occasionally heard voices stating that "coverages are features". Now, true, but what does this help us in practice? Some people start encoding coverages in GML which I find not the best of all possible ideas when it comes to a generic (n-D) solution. Now you are AFAIK the first one to state that the inverse holds as well: "features are coverages, sort of". A remarkable point in history! ;-) Seriously, I believe that when it comes to border cases there is a degree of freedom to the data designer, which is not bad. So I feel that WFS and WCS nicely coexist, augmented by the third companion, metadata served by a CS-W. SOS (and SWE in general) IMHO sometimes has been a little unimpressed by existing standards (disclaimer: my private opinion!). Now we have some work to do in harmonizing, which will be a big topic at our next TC meeting in December, involving several WGs. cheers, Peter Jon Blower wrote:
Hi all, Can I come at this from a slightly different angle - what is considered *out* of scope for WCS? It is often said that a successful business must be very clear about what it *doesn't* do (I saw it in a Dilbert cartoon so it must be true), and the same is surely true for standards. Back in the early days, WCS only dealt with 2-D rasters, reflecting its origins in satellite imagery. Although this was limiting (and highly unsuitable for some communities), it was at least implementable. Independently-developed clients and servers could interoperate. However, now it seems that nothing is out of scope for WCS. Furthermore, the ISO19123 Coverage model itself is also extremely general and it seems that pretty much any data can now be described as a Coverage, including what we might once have described as a "feature" (non-raster data). This considerably blurs the distinction between WCS, WFS and SOS. So, can people help me to understand what kinds of data would *not* be considered suitable for a WCS approach? Cheers, Jon
galeon archives: