SRS handling questions

[Silke Reimer]

On Thu, Mar 25, 2004 at 07:45:05PM +0100, Martin Schulze wrote:
> Silke Reimer wrote:
> > > My interpretation is that a layer is available for the SRS it
> > > announces in the <BoundingBox> elements (and llbbox) with the
> > > respective minimum enclosing rectangle as specified.
> > 
> > In my opinion the announcement of SRS is done in the SRS-element not
> > in the BoundingBox-element (see above).
> 
> > You quote:
> >
> >      [..] A server which has the ability to transform data to different
> >      SRSes may choose not to provide an explicit BoundingBox for every
> >      possible SRS available for each Layer.
> >
> >      The server should provide BoundingBox information for a tleast the
> >      native SRS of the Layer.
> >
> > Hence you can announce SRSes without giving a BoundingBox for a
> > specific SRS. I interpretate this as: The announcement is done in
> > the SRS element while you *can* provide a BoundingBox. 
> 
> *blush* You are correct.  I don't know how I managed to overlook this
> during reading an typing.
> 
> Doesn't this pose a problem with non-existent knowledge in the client
> about the known rectangle in that particular projection?  The server
> can cope with it without a problem.

Hm, this depends on the ability of the client to reproject by
himself. Obviously there must be at least on definition of a
BoundingBox or LatLonBoundingBox. In addition  I remember that one
of those BoundingBoxes (I think it was LatLonBoundingBox but this
has be checked in the spec) must contain the union of all rectangles
which is covered by the projections. Thus there is a definition of
an overall BoundingBox.

I am not sure what happens if no BoundingBox is provided. I would
assume that the specific projection than covers the union of all
BoundingBoxes thus the rectangle defined by the LatLonBoundingBox.

And then: If the client requests a wrong BoundingBox the worse is
that he gets an empty image. This is not optimal but doesn't crash
the client. I think the greatest problem is the "zoom to full
extent" functionality of the client.

	Silke

-- 
Silke Reimer

Intevation GmbH                      http://intevation.de/
FreeGIS                                http://freegis.org/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://www.intevation.de/pipermail/thuban-devel/attachments/20040326/16b08934/attachment.bin