OK, I got break processing figured out :)

Now, how do I deal with the case where I'm given a table with an X field and an
arbitrary number of (named) Y fields? I think I just need some hints to get
started.

For example, I have a table of 5-minute system load averages from a set of Linux
systems, and the size of this set may/will vary for each different report run:

tPosix,tString,device1,device2,device3,device4,...etc.
1264527900,"2010-01-26 09:45:00
PST",1.53e+02,1.35e+02,1.22e+02,1.39e+02,2.34e+02,

I need to create, first, a scatter plot and include a legend giving the column
(device) names.

What are the key words/concepts I need to look into? I *suppose* I could do
some trivial pre-processing to set an environment variable to the table width,
if necessary.

Thanks.