Nitro wrote: > just want to chime in and say that fc2 does hit-testing differently, > similar to the way you patched the fc1 code (bounding box as the first > coarse test). It also uses the spatial query facility (that's what > hit-testing is called in fc2) to cull away objects that are not within the > viewport.
FC-1 does that too -- you may have noticed that it's a lot faster to draw when you are zoomed in and there is less to draw. > Right now the culling and hit-testing is done in a linear fashion, by > iterating over all objects in a list. However, there's already a stub > implementation in the rtree.py file (which does things in the linear > fashion right now). It is likely not difficult to integrate it with > http://pypi.python.org/pypi/Rtree (C library for rtrees) to perform > hit-testing and culling in a much faster way. > > Not sure if this helps you, but I wanted to mention it before you do a lot > of extra work... Thanks for the note, Matt. It looks like we'll be hiring someone soon to build a tool based on FloatCanvas -- that should allow us to dig into the code again, and hopefully FC2 working right on GTK and OS-X, and then see where it takes us... -Chris -- Christopher Barker, Ph.D. Oceanographer Emergency Response Division NOAA/NOS/OR&R (206) 526-6959 voice 7600 Sand Point Way NE (206) 526-6329 fax Seattle, WA 98115 (206) 526-6317 main reception [email protected] _______________________________________________ FloatCanvas mailing list [email protected] http://mail.mithis.com/cgi-bin/mailman/listinfo/floatcanvas
