issues with 'orbits'



Hi Sheldon,
thanks for your report.

On Qui, 2009-10-22 at 18:47 -0400, Sheldon Newhouse wrote:
> If one adds the option as in
> orbits(r*x*(1-x),.2,100,200,[r,3,4,.1],[plot_format, openmath]);
> 
> then the plot  is produced almost 2 or 3 seconds after the words "Graph 
> passed to plot2d" 
> 
> So, it seems that the interface to gnuplot has some problems.
I experienced a similar problem in Linux using GCL. It looked like a
gnuplot problem but it did not show up when I switched from GCL to other
Lisp flavors. It sounds strange but it might be GCL issue.

> 2. In linux (Fedora Core 8) this problem does not occur.  The plot is 
> drawn soon after the words "Graph passed to plot2d"  appear.
I guess you are not using GCL in Fedora. If you can, please try it with GCL.

> 3. In linux ( don't know about WinXP) there is a problem with the 
> command 'orbits' if the package 'dynamics' is already loaded:
> Maxima version: 5.19.2
> lisp-implementation-type: CMU Common Lisp
> lisp-implementation-version: 19d Fedora release 5.fc8 (19D)
> (%i5) load(dynamics);
> (%o5) "/usr/local/share/maxima/5.19.2/share/dynamics/dynamics.mac"
> (%i6) orbits(r*x*(1-x),.2,100,200,[r,3,4,.1]);
> Graph passed to plot2d...
> Improper name or value in functional position:
> errexp1
> #0: 
> orbits(f=r*(1-x)*x,y0=0.2,n1=100,n2=200,domain=[r,3,4,0.1],options=[])(dynamics.mac 
> line 367)
I cannot reproduce that error in Ubuntu with either Clisp or Clozure
Common Lisp (I do not have CMUCL at hand). It might be a CMUCL issue; in
any case, you no longer need to use load("dynamics")

For the record, here are the times that command takes in the same
machine (Intel Quad CPU Q6600) using three different Lisp flavors:

Clisp:   31.4400 seconds, using 608.044 MB.
SBCL:    11.3200 seconds, using 535.312 MB.
Opencml:  7.3800 seconds (no info about memory usage).

> 4. In general, replotting in the 'openmath' window has many bugs.  Try 
> zooming, unzooming, or resizing to the whole screen and replotting for 
> instance.

Yes I'm aware of those bugs, although it would be nice to have them
registered in the bug-tacking system. Moving the plots also shows lots
of bugs. I have decided to make a new version of plot2d.tcl from scratch
which is easier than finding those bugs. Hopefully, I will have
something to show for version 5.20.

Regards,
Jaime