Jon Blower wrote:
Hi John and Galeon folk,
Excellent points. Just a couple of comments:
So its important to acknowledge that WCS can bo things that opendap/netcdf cant.
This is true, although this alone isn't a justification to spend huge
effort designing and adopting a whole new protocol. Modifying
existing protocols is another valid approach. Also, users don't type
in WCS or OPeNDAP URLs manually - its done with a tool. Of course,
this assumes we have CF-aware tools for every language, which we
don't, so pushing this to the server is attractive.
The protocol, though, isn't completely new. It is, however, in
transition. I think the point, though, is in John C.'s statement: The
WCS standard allows things that NetCDF doesn't, and vice versa. Pushing
this to the server *does* make sense.
11. So can we use WCS to deliver data? Id say the current answer is "yes, but
without clients, who cares?".
Agreed - and I'm not convinced that the major GIS vendors are going to
build good-quality clients in the foreseeable future. I've talked to
a few GIS vendors informally and although they "buy" the WMS idea,
they are much less convinced about WCS, WFS and so forth. They pay
lip service to these standards to give the appearance of being "open"
and "up with the game" but do not (yet) seem very interested in
committing. This is probably partly due to conservatism and partly
scepticism but also because they have their own proprietary data
models and services that they want to keep pushing. Have others had
different impressions of attitudes of GIS vendors?
Adoption in the course of a software development and acceptance process
takes time, and there is, I'm sure, a bit of the old closed-source
culture to contend with. However, the evolving nature of the WCS
standard is such that it may well take time for commercial vendors to
incorporate something. I expect FOSS to develop a decent reference
implementation and release it first, although I'm not going to guess who.
Note that the various OWS efforts have yielded commercially undertaken
WCS servers... and clients... and that in the US, WCS is being adopted
by the National Geospatial Agency.
I think the solution to the client problem is to keep things simple
and make it easy for client developers to do the right thing. Then we
can make at least some progress. Complexity can come later following
buy-in from users and software vendors.
Always a good view.
gerry