The problem with folding it into project is that it has a few map
graphics. Maybe just the interface and renderers?
On 6-Nov-08, at 9:22 AM, Jody Garnett wrote:
You could also fold MapGraphic into project - I do want to collapse
the plugins together and mapgraphic does not add too much value on
its own - there may be the danger of breaking the extension point
however (cannot remember if the plugin id is involved in the
extension point name or not?).
However ...
Emily I am not sure anybody needs to know - the MapGraphic renderer
implementation can do whatever it wants right. You can do what you
need to without violating information hiding; nobody needs to know
you are rendering to a screen sized image outside of the
MapGraphicRenderer... remember the only thing that renderer needs to
do is correctly play the game with the renderer composition
correctly (if a big image is stored behind the sceen nobody needs to
find out...).
Jody
Jesse Eichar wrote:
You can detect by seeing if the IGeoResource resolves to
MapGraphic. You should almost never do instanceof checks in uDig.
But I think I understand the issue. Project does not depend on
MapGraphic so how can you do the detection right? Honestly not
sure. What options do we have?
1. Make a marker interface ?
2. Add something to the render metrics ?
3. Have a extension point where one can declare ways to handle
renderers?
Any other options?
Jesse
On 5-Nov-08, at 8:53 PM, Emily Gouge wrote:
Hmm I think I get what you are trying to say here.
I would need a single list of mapgraphics layers that are rendered
to a "screen tile". Then when I go to compose my tiles I can
extract the section of the mapgraphic image that corresponds to
the section of the tile I am composing. Hmm sounds like fun to
implement!
One question - If I want to start treating MapGraphics as
"special" renderers (so I can do different things based on the
type of renderer) I need a way to distinguish between regular
renderers and mapgraphic renderers. All the mapgraphic classes
are in a different package so I can't simply to an "instanceof
MapGraphic" test and I know it's not the correct thing to add the
MapGraphic package as a dependency; so how do I go about
determining if a renderer is a "special" renderer (in this case a
mapgraphic renderer).
Thanks
Emily
Jody Garnett wrote:
Jesse the "tiling" system would still need to suck out a small
tile worth of the main mapgraphic image when composing a "tile"
for display ... unless the one specific case where the mapgraphic
is "ontop" of the data layers that are all being tiled up.
Emily I would suggest having the mapgraphic write into an image
the size of the screen; and giving the "tiled renderer" the
ability to select out a rectangle from this large image for use
when "composing" the BufferedImage for an individual Tile...
Not sure if I explained that very well...
Jody
Jesse Eichar wrote:
IMHO I think that MapGraphics should not be part of the tiling.
The reasoning is that they are supposed to be light-weight and
fast. They should normally be fast enough without tiling. And
often are independent of zoom so how would you know when to
cache and when not to.
Jesse
On 4-Nov-08, at 11:49 PM, Emily Gouge wrote:
All,
I'm trying to clean up the the tiled rendering system I
worked. One of the things I'm currently looking at is trying
to get mapgraphics to work nicely with it.
My understanding is that currently mapgraphics pay attention to
the bounds of the screen. So in my tiled rendering system they
are paying attention to the bounds of the screen and using that
information to draw onto the tiles. These tiles don't match up
with the screen so as a result you get some fun results:
* legends, north arrows, etc on every tile.
* grids only partially filled in (if then render tiles are
bigger than the screen only what would fit on the screen is
filled in on the render tile).
So how to fix this?
One option is to not use the tiled rendering system for
mapgraphics.
As a result mapgraphics would always render on top of the tiles
and you couldn't have map graphics in between layers or at the
bottom of the stack. My feeling is that this is undesirable
behaviour.
Another option may be to have the mapgraphics pay attention to
the render bounds as well as the screen bounds and size; but I
think this will cause a lot of unnecessary complexity in the
mapgraphics.
Does anybody have an other ideas on how to fix this problem?
Thanks,
Emily
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
