Ultimately I need to place some 2-D plots in my LaTex publication. I played with PGPLOT for a while, but I wasn't able to figure out how to make the color/linestyle of the grid different from the axes and labels (I strongly prefer a light gray grid color with black axes and tic labels). Also I really want fine control over the tic locations which doesn't seem possible at least with the high-level PGPLOT commands. I suspect it's doable with low-level type commands; but my efforts are slow going in figuring these out.
So I changed up my workflow plan to proceed with PDL/PGPLOT for analysis and then export my final datasets to text files for plotting in Gnuplot -> PSTricks -> LaTeX. This strategy is solid; but slightly above my laziness threshold. This means I've comitted all my resources to figuring out how to skip PGPLOT and do both my analysis and publication graphics with Gnuplot. I've had pretty good success using the GnuplotIF perl module. Of course I have to flatten piddles into perl lists before shipping them off to Gnuplot using list(). The problem is sporadic data corruption. Sometimes all the data gets through, sometimes not. Large datasets are a disaster. For example, 100 datapoints may get through uncorrupted one time. Then if I resend the same data data a few times, I'm likely to get a couple of points dropped. If I send 5000 datapoints, the plot is severely messed up. The workaround that I've found is to dump the data to a temporary text file with wcols() and then use the Gnuplot interface to open that text file back up. It feels a little sloppy but that's what I'm going to stick with for now. I welcome comments on anything. The code in GnuplotIF.pm looks OK to my eyes and the glitch occurrence is not consistent. I'm not sure how to debug the problem further but if this is at all interesting to someone more code savvy than myself, feel free to let me know what you find. Also, I'd be interested to know if the problem even exists anywhere else besides my computer and/or os platform (Activestate/Windows XP environment). The problem should be reproducible with this code: http://pastebin.com/m809dc15 For me, setting $npoints to 100 will drop data only occasionally. Bumping the value up to 500 or 1000 messes things up pretty quickly. And to give you an idea what I'm going for, see here: http://img520.imageshack.us/img520/395/sinctest.png If I set the output to PSTricks in Gnuplot and send this graph through LaTeX I can get a nice publication ready rendering. I'm also interested in any thoughts on accomplishing this sort of thing with PGPLOT. Thanks! --Art
_______________________________________________ Perldl mailing list [email protected] http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
