Progress Report for the week of January 20, 1998

Accomplishments:

This week we spent Saturday in the lab working out some bugs in our code. We were able to get JViz working on Solaris. We had to make one change in the code, though. We had to remove one of the test for a mouse event. I guess Solaris has different mouse clicking events than Windows. Other than that, we didn't have to change anything. We also got it to work on the Alphas without modification. Performance was comperable to that on my NT machine. We did find some disturbing bugs/happenings when working on crane from the X-tubes in the CCC. For some strange reason, some of the graphs would not display on one machine, while running the exact same thing from another station had no problems whatsoever. the problem on the original machine did not go away until completely logging out and logging back in. Very strange. Will made good progress on the scatter plot graphs. He has put the labels on and also fixed the problem with the brush. It is looking pretty good. Jason also made good progress on the glyph graph, but it is still a little buggy at times. He has implemented the activation and deactivation of dimensions successfully and clicking on the glyphs changes the brush, but he is still working on getting the radial lines in. Also, on large data sets, we ran into an out of memory error. The images that he was creating to draw to was too big to fit into memory. We are working on getting around that problem. We will run into that problem on every graph, so we have to come up with an intelligent solution to it for all of them. Matt spent considerable time trying to get the scroll pane to show scroll bars when using the latest version of swing. It has worked fine in all previous versions of swing, but not the most current. I tried looking on the newsgroups, at Sun, asking people I know, and I seem to be the only one with this exact problem. The final release version of the library is due out in a couple weeks. For now we will have to use 0.6.1 instead of 0.7. The Color Manager has proven to be quite a bothersome piece. We were hoping to use the built in color selector class, but it didn't provide the functionality and customization we needed, so we have decided to implement our own version. It will look very similar to the one currently used in XmdvTool. We found what was causing the bug in the brush. It was actually in the data set. both max and min were set to be 0. What should we do in this case? Finally, we got ahold of the "runnit" script and batch file that Sun uses for their demo programs. As long as you have the specified environment variable set (path to root of swing directory), then all you have to do is type runnit to run the program. We've tested the batch file version, but not the shell script yet, but there is not reason to think it won't work.

Questions:

To do List:
BACK to MQP homepage