Maxima-Gnuplot interface



James Amundson ÐÉÛÅÔ:

> On Sat, 2004-04-17 at 12:15, Vadim V. Zhytnikov wrote:
> 
> 
>>I've tried new gnuplot plotting on both Linux and Windows XP.  Works 
>>fine especially in conjunction with gnuplot 4.0.0 released yesterday.
> 
> 
> I am particularly pleased to hear that the new interface is working on
> windows. The only testing I did there was months ago, when the interface
> was still in the concept-only stage.
> 
> 
>>I can only confirm two glitches reported by Valery.
>>
>>1) plot2d(cot(x),[x,-2*%pi,2*%pi]);
>>
>>produces Zero-Divisor lisp error on GCL (clisp, sbcl and cmucl are OK).
> 
> 
> It would be nice if we could understand this problem.
> 
> 

Here is a bunch of plot problems.  They are not directly related
to new plotting machinery but maybe some of them may be fixed before 
release.

PLOT2D:

(1)
plot2d(1,[x,0,1]) or plot2d(x,[x,0,1])
- lisp error (not intercepted in GCL). plot2d don't want
to plot atomic expression ([1] or [x] are OK).

(2) plotting with unknown function of variable:
plot2d(foo(),[x,0,1])
plot2d(foo(x),[x,0,1])
plot2d(foo^2,[x,0,1])
- No maxima or lisp error but gnuplot error instead. Maxima tries to 
plot empty plot.

(3) plotting with unknown function of variable:
plot2d([foo],[x,0,1])
- obscure error message: ERROR while printing error message.

(4) plotting with unknown function of variable:
plot2d(foo*sin(x),[x,0,1])
- no errors, empty plot.

(5) singular functions:
plot2d(cot(x),[x,-1,1])
- lisp error, on GCL it remains non intercepted and
dumps user to lisp error prompt.

(6)
plot2d(1/x,[x,-1,1])
- strange plot on sbcl.

PLOT3D:

(7)
plot3d(foo,[x,0,1],[y,0,1])
- error like in (3).

(8)
plot3d(foo(x,y),[x,0,1],[y,0,1])
- lisp error. On gcl dumps user to lisp prompt.

(9)
plot3d(cot(x*y),[x,0,1],[y,0,1])
plot3d(1/x,[x,0,1],[y,0,1])
- lisp or gnuplot error.

Mostly we have here two issues:

1. GCL specific problems - some lisp errors.

2. Problems with singular or not well defined functions.
If expression has several singularities within plotting
range they probably cold be safely ignored.  But how to
distinguish this from the expression which just can't be
plotted due to unknown function or variable?





-- 
      Vadim V. Zhytnikov

       <vvzhy@mail.ru>
      <vvzhy@netorn.ru>