Hi,
Larry, you really know how to use a search engine :-) .
You're right, these experiments did not involve points and I really have
no idea about how to display a point faster. The idea in these mails was
that a single pixel is enough to display all the features included in a
rectangle which is less than a pixel size, and that quadtree or rtree
structures can give us such an information before features to display
are entirely parsed or decimated.
Michaël
Larry Becker a écrit :
Hi Michaël,
I was referring to this note in Java2DConverter:
// [NOTE] I tested a short-circuit here to return a 1 pixel
representation of
// geometries having an envelope of less than 1/2 pixel, but I
get only an ugly
// render for a small performance improvement [mmichaud -
2007-05-23]
and your post of March 9, 2007:
"PS : my experience to use a spatial index to accelerate graphic display
when there are many features to display :
I made some experiences in the past using a quad tree and a technique
which is quite close to yours :
as soon as I reach a cell of the spatial index which is less than one
pixel
size (or 1/2), I just displayed the pixel instead of reading all the
features."
But perhaps, your experiments didn't involve points.
regards,
Larry
_______________________________________________
jump-users mailing list
[email protected]
http://lists.refractions.net/mailman/listinfo/jump-users
