This function does not depend on y. If you use a failure
on this pathological case as a reason
to abandon the lisp plotting functionality, then what
would you do when some nonsense input to gnuplot makes
it behave oddly... abandon it too? :)
The commercial macsyma has no problem with this,
incidentally.
There are lots of other plotting programs that I think
produce better output than gnuplot (but isn't it gplot?),
though maybe not free. Certainly there has to be
a community of scientific visualization experts out there.
I personally think that getting plotting right is
a hugely complicated endeavor. One possibility is to
look at Mathematica's plot package design and clone it.
Or commercial Macsyma's plot package.. which in some
respects is better than Mathematica, but maybe too
complex, also. Or at Maple, Mupad, etc..
In some cases the key to a good plotting program is
fast computation of the function being plotted, hence
translation into lisp, or into binary code can help. Also
stuff like fast computation on a grid by differencing,
as done by Zima and Wang.
RJF
willisb@unk.edu wrote:
>
>
> I don't read sloop code, but I speculate that draw2d (defined in plot.lisp)
> does a adaptive evaluation of the function -- presumably where the function
> changes rapidly, it is evaluated on a finer (and finer) grid. Does
> plot2d (or friends) get stuck in a loop taking too fine a grid?
>
> Actually, I wonder if we'd be better off developing a better gnuplot
> interface
> for Maxima than improving plot.lisp -- Do we have the resources to write
> really good graphical code from scratch? The mygnuplot function
> [see describe("system") ] could be improved without a great deal of
> work -- I've hacked on it a bit so that it works better for me.
>
> Or maybe somebody could write an interface to something other than gnuplot?
>
>
> Barton
>
>
>
>>>>>>"Rich" == Rich Drewes <drewes@interstice.com> writes:
>>>>>>
>
> Rich> Consider this Maxima input:
>
> Rich> f(x):=(7-500*x)/(25 * x);
> Rich> plot2d(f(x), [x, .008, .012], [y, 2, 16]);
>
> Rich> The plot takes upwards of 20 minutes to appear on a reasonably
> fast
> Rich> system. Something pathological seems to be going on with
> certain
> Rich> domains for the plot of many simple functions. Also, when the
> plot
> Rich> finally does appear, the y axis doesn't seem to be present (even
> when
> Rich> the plot is moved around the window) and the x axis doesn't have
> Rich> reasonable tick marks.
>
>
>
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>