Bart van den Eijnden (OSGIS) wrote:
Hi Jody,

Right now I am testing against a UMN Mapserver. But I can't see any
requests in our access log wrt GetLegendGraphic at all.
Actually now that I think of it I think we do use the legendURL, it is just in my experience that this
usually points to a GetLegendGraphics request.
LegendURL is in the WMS GetCapabilities response. But UMN Mapserver by
default advertizes 18x12 pixels, whereas the generated legends can be much
larger than that (300 pixels eg). In the WMS 1.3 spec this has been fixed
by making width/height optional for LegendURL. So the best (most reliable)
thing a client can do is just inspect the image to determine the actual
size, but I don't know if uDIG could handle that.
Udig does inspect the returned image, and resizes or crops as best it can guess. If you have
acess to the codebase I can walk you through the relative classes via IRC>
So but basically the legend will only be displayed in the Layer view?
There is no way to get a larger legend in uDIG? What happens then to
layers like soil maps with 250 classes and big legends?
uDig layers view is not a legend, the focus is on zorder and layer management rather then layer explanation. As a consequence we are only using the legend graphics to "best represent" the layer in uDig.

Note: I did try having the larger graphics in uDig and the result was an inconsistent mess :-(

A couple ideas:
- presenting the larger image as a tooltip...
- use of the Legend MapGraphic (you can place on a map for viewing or printing, and it could easily make better use of larger legend graphics)

As for large layers like "soil maps" I have no good suggestions, the WMS specification does not provide enough information for us to reliably discover categorization information - a lack we really would like to address so we can make use of WMS SLD with custom SLD files.

Where would you like to go from here?

Cheers,
Jody

_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel

Reply via email to