[Fwd: [OSGeo-Standards] WMS-C and Capabilities]

[Christopher Schmidt]

On Thu, Aug 28, 2008 at 02:03:59PM +1000, Mark Leslie wrote:
> Jody Garnett wrote:
> 
> >Christopher Schmidt wrote:
> >> On Tue, Aug 26, 2008 at 11:53:33AM -0700, Jody Garnett wrote:
> >>
> >>> This discussion is fascinating to me as it does not align with my
> >>> expectations ... my expectation is that this
> >>> WMS-C documents some additional vendor specific capabilities that are
> >>> added to an existing WMS (much like SLD support is added to an existing
> >>> WMS).
> >>>
> >>> As such I am expecting to see a WMS with vendor specific capabilities
> >>> letting me know that some TileSets are available; and that if I call
> the
> >>> existing GetMap operation in just the right way I can make use of it.
> >>>
> >>> So I was not expecting to see a standalone TileCache.
> >>>
> >> Right. I'm not aware of any caching tile server which also serves
> >> non-cached data with regular WMS requests, but the WMS spec was designed
> >> to allow that.
> >>
> >I was more thinking for a WMS implementation that had some caching
> >support built in. Sounds like we have set up the convention to allow for
> >this; but lack a proof of concept.
> >
> >However the point is made - Andrea is correctly describing the existing
> >implementations. While this is not what I expected I can certainly roll
> >with it; now if only there was a good way to tell where the source WMS
> >server is :-P
> >
> >Thanks for answering my questions everyone.
> >Jody
> 
> There's another (potential) standard to consider, which is the OpenGIS
> Web Map Tiling Service Discussion Paper.  From what I can gather, this
> is the evolution of the wmsc proposal, but they have made the decision
> to clearly separate it from the WMS spec, GetTile instead of GetMap for
> example.  It's not a spec yet, but OWS-6 is looking to change that from
> the sounds of it.
> I don't have a link to the paper, but the number is OGC 07-057r4.

The paper isn't published; no one who isn't an OGC member can get access
to it.

As usual, the OGC is going about creating standards in a vaccuum, hiding
drafts from potential implementors. 

(TileCache has had WMTS support for 9 months now, but only because I got
an illegal copy of the document from someone who does have OGC access.
I've been waiting for the document to be published ever since so I could
comment on it, since there are a number of supremely flawed design
decisions, imho.) 

Regards,
-- 
Christopher Schmidt
MetaCarta