Due to the current gap in continued funding from the U.S. National Science Foundation (NSF), the NSF Unidata Program Center has temporarily paused most operations. See NSF Unidata Pause in Most Operations for details.
Hi Bill, > There are two stages in "rendering" a DataReference: > 1. Transforming the Data to a scene graph, and > 2. Java3D's rendering of the scene graph to the JPanel. > When you call the setData() method of selected > DataReferences, only those DataReferences should be > re-transformed, unless you are constantly re-autoscaling > (which can happen if there are ScalarMaps whose RealType > values are always missing). I'm not doing any re-autoscaling. But perhaps the problem is my datastructure. I have about 1080 GeoLocations to manage, e.g. 1080 FlatFields each of them holding data for two Longitude-Latitude-Points. So each FlatField represents a line. I don't know how fast VisAD can handle such data. Is there a way to improve my datastructure? I need to draw every GeoLocation as single line not as polyline. So a MathType like (index -> (longitude, latitude)) won't work. Therefore I've chosen to manage my data as (index -> (longitude -> latitude)) as you can see from the source of the former mail. > Re-transforming is usually a lot slower than re-rendering > by Java3D, so if your application is slow then your map > outlines may be re-transformed. Are you calling setData() > for your map outlines, when you are updating your other > DataReferences? Or are you always re-autoscaling? Nothing of both. Thanks for your help Bill, Mathias
visad
archives: