> Looking around I see that "JVX" stands for Java View geometry file. JVXL > is > probably better?
another random idea ... .jso file > René, OK, I understand what the code is doing now. I calculate > that inclusion of interpolation smoothing still allows for the > inside/outside idea, but along with that, an accounting is > necessary of the fractional distance along each > edge of the isosurface "voxel" to offset the surface point. Correct. > Assuming that that > fractional distance can be assigned a number 0-63, I calculate a new > compression > ratio of 500:1 instead of 1000:1 with straight 1-byte-per-edge storage. Excellent. > This > changes to about 600:1 if the range can be simplified to 0-7 and these are > stored two values per ascii character -- (probably not worth it to > compress that > much). > See > http://www.stolaf.edu/people/hansonr/jmol/test/proto/ch3cl.jvxl, > which is now about 7400 bytes and includes 0-7 range interpolation data. > > A script command describing an orbital might be as long as 4000 > characters. My > idea would be to put this in a script file and just call it. Or, maybe we > don't > need the inline idea, and we can just load the orbital. I personally don't see much use for the 'inline' stuff, but with all of your JavaScript work you have gotten a lot of mileage out of it. > An interesting thing about cube (jvxl) files is that once read, the > "orbital" > can be rotated and scaled to any size and any position, so this might just > be > the way to express generic hybrid orbitals for display at any atom. A bit > of > calculation and some good syntax should, for example, allow: > > > orbital sp3 1.0 (oxygen and connected(2)) > > orbital p 0.8 (carbon and connected(3)) > > (where the number is the scale) all based on a single hardwired (or > user-loaded?) isosurface representing a single tear-drop-shaped orbital > lobe. > > The idea here would not be to be exact in every detail -- just provide > enough of > a surface for depicting the essence of the orbital in a compact format. Bob, I'm thinking about the fact that you are pre-computing ... Q: Is there a reasonable range of scalar data values that show up in the data files? The SqueezeCube program that you asked about was saving space by taking small values (< 0.001) and 0.000000 values and turning them into '0' Your idea of 'run-length' compressing these things is excellent. It makes me wonder if we can get most of the savings by eliminating the 0's, while retaining the flexibility of having the scalar values so that one can calculate the actual surface on the fly. ** 2 minutes later ** In thinking about your compression, I bet it would be interesting to try to apply some type of image compression to this data ... like jpeg. These compression formats have the advantage of providing approximations to interpolated values. Going a step further, we really would want to apply some type of mpeg compression. Think of it this way ... layers along the z-dimension are frames in a movie. The image does not have much change when going from frame-to-frame. Within a frame the image does not change much when going from pixel-to-pixel. Miguel ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642 _______________________________________________ Jmol-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jmol-developers
