I am working through the design of a disp[lay for a CAD/GIS program
written
in Java. I haven't decided for sure if I will use Swing or SWT for the
GUI yet.

This program I am working on will have to display a large number of 2D
geometric shapes. I'm want to make the rendering process as quick and
efficient as possible. I had a couple of questions about how I can do
this:

Let's say for the sake of this discussion that the area I will use to
display my map, or the canvas, will be 600 by 300 pixels. Would it be
possible to render the shapes that would fall within an area 20% larger
than
this to allow for quicker panning and zooming?

I'm trying to avoid a situation where everytime the user pans the
display a
minor amount that the whole screen needs to be rendered. In the above
situation I would have that 20% buffer to allow some limited space for
panning.

Is this method possible? Any ideas or suggestions on how I can learn
the
techniques I need to implement it?

Also, my program will allow the user to select shapes on the display
using
various interactions with the mouse, and will subsequently allow them
to
move or modify those geometries. Is it possible to only render the
portion
of the canvas that needs to be repainted, instead of repainting the
entire
canvas everytime a single shape is modified? Would I do this by
splitting my
display into a grid of different images, or can I just pass the
rectangular
bounds of the new area that needs to be rendered?

Is this method of partial rendering possible? Any ideas or suggestions
on
how I can learn the techniques I need to implement it?

I am interested in learning how this might be done in both Swing and
SWT. The way each widget toolkit handles this area of graphics display
may be the diciding factor in which is chosen for the program.

Thanks for the help.

Scott Huey