Jean-Marc Valin wrote: > > As Brad is saying, there are currently 120 implemented node types, so doing an > icon for each is not feasable in short-term. What could be done is having an > icon for some of the most common nodes (Add, Bub, Mul, AND, OR, Constant, Load, > Save, ...) and then we can have "category" icons like "DSP", "neural nets", > "Vector quantization", ... There are about 15 categories. This is what I meant. I'm aware that there are 100's of actual nodes and many more to come. What I meant by "types" is "categories", classes, etc. 15 or so may be doable. > Also, when building a > subnet, we should have the option to assign an icon to it. Right. Since a subnet can be made to do anything, subnet icons should be changable. > I'd like the whole thing to look a bit like Simulink. I'm not sure I've heard of Simulink. Do you have a URL? Cheers. Jeff -- +----------------------------------+ | J.W. Bizzaro | | | | http://bioinformatics.org/~jeff/ | | | | BIOINFORMATICS.ORG | | The Open Lab | | | | http://bioinformatics.org/ | +----------------------------------+