On Sat, 2004-01-31 at 22:15, Jay Belanger wrote:
> The nice thing about mgnuplot is that the user can interact with it
> after the graph is plotted.
> It'd be much nicer if such things could be done directly from maxima,
> but common lisp doesn't seem to have any subprocess abilities. (Is that
> right?)
> Without subprocesses, the maxout.gnuplot file could be kept, and after
> a Maxima plotting command, it could be modified and gnuplot restarted,
> but that seems unpleasant at best.
I was originally going to avoid subprocesses. I think talking to gnuplot
via pipes is a better solution, but I didn't want to take the time to do
it right now. However, as Ray has pointed out, clocc already contains a
module to talk to gnuplot via pipes. I knew that once, but had forgotten
it. I haven't had time to play with the clocc implementation yet.
In the worst-case scenario, maxima could still call gnuplot, let the
user enter more options based on what he/she sees, then call gnuplot
again with the revised input.
> (What constitutes a proof-of-proof-of-concept, by the way?)
I had a version of maxima that called gnuplot to create a graph and
leave the graph window open. That was the part I was unsure about. I
only gave it (proof-of)^2 because (1) I had to modify it by hand when I
moved between windows and Linux and (2) I didn't generate a
maxout.gnuplot file -- I just used an old one copied into the current
working directory.
--Jim