I think this is a very interesting idea. Has
anyone looked at performance of Gef & Draw2D? I wrote a simple program
that connected a bunch of boxes with Polylines (using draw2d) and put a mouse
listener on each box (to handle drags). After about 10K nodes this got
very very slow. I mentioned this to a friend of mine who works at Nvidia
and he said he could do the same thing (in C obviously) and get about 10Mil
boxes (tapping into the H/W acceleration etc...). I was wondering if we
could hit something like this (or even 100K nodes) using OpenGL or other H/W
accelerated drawing surfaces. (Or maybe Draw2D wasn't ever designed for
this)
Just a thought.
cheers,
Ian
----- Original Message -----
Sent: Monday, June 28, 2004 12:43
AM
Subject: RE: [gef-dev] Looking for Next
Release (3.1?) Requirements
Oh and I forget something that migth be worthwhile: A
bridge between Draw2d and OpenGL. Currently Draw2D renders to a SWT GC using
SWT images, doesn't it? Maybe it's possible to implement something like an
OpenGL based LWS. Not sure, how usable the SWT OpenGL interface/plug-in
is.
Cu, Gunnar
With 3.0 almost out the door, we
will be prioritizing features and bugs for the next major release of GEF
(and also the service release 3.0.1). If you have requirements, bring
them up here, or if we get too many reponses, bugzillas are probably the
more appropriate forum.
Some
ideas so far: - New examples
- Graphical Text support - Snap bendpoint to grid - New draw2d layout (table?) - Improvements to the DirectedGraphLayout
(constraints, optimizations)
Please search for bugzillas before opening a new one. Also,
describe what you are doing (the problem) first, and then suggest the
solution.
-Randy
Hudson
|