[OSGeo-Standards] WMS-C and Capabilities

Jody Garnett jgarnett at refractions.net
Tue Aug 26 15:04:51 EDT 2008

Andrea Aime wrote:
> I did not explain myself. I do not mean to say the capabilities document
> should be partial or anything, I just mean the client cannot expect
> a tiling server to answer any kind of request like a WMS would do.
Thanks Andrea that communicates very clearly.
> The capabilities document is an extension all right, but the behaviour
> is a restriction, only certain GetMap requests will be answered, no?
I was personally viewing it as an addition; the promiss is that only 
certain GetMap requests would be cached. I can work with this either way 
- I think the wiki page documents the option of throwing an error if 
tiled=true is not provided for example?
> Put in other words, the caps document is based on WMS, but the tiling  
> server is not a valid WMS server, right?
Perhaps we can use the word "general purpose" rather than valid?
If my understanding of tiled=true being required is correct; then I will 
agree - WMSC is not a general purpose WMS. It has additional 
restrictions that must be met - and thus can not satisfy any GetMap request.

The question for me then becomes (as a client writer) how can I tell a 
general purpose WMS that supports caching  (ie has TileSet information) 
from a WMS-C that only serves entries that align with TileSets?


More information about the Standards mailing list