plot2d failed to plot too many discrete points

Richard Fateman <fateman at eecs.berkeley.edu> writes:

> Another idea
>
> On 5/13/2012 2:09 PM, Richard Fateman wrote:
>> On 5/13/2012 7:43 AM, XeCycle wrote:
>>> Richard Fateman<fateman at eecs.berkeley.edu>  writes:
>>>
>>> [...]
>>>
>>>> I have  fairly nice computer display, and it has about 1.6
>>>> million discrete pixels.
>>>> Plotting 2 million discrete non-overlapping points on it would
>>>> not be possible.
>>>> Perhaps you would find difficulties regardless of the plotting program.
>>>> RJF
>>> You're missing the point.  This many numbers are needed for my
>>> computation, and will be further processed; now I want to have an
>>> impression on how it looks like, so I tried to plot it.
>
> If you want to know what it looks like, how about plotting only every 
> 1000th point.
> That still gives you several thousand points plotted, and that may give 
> you an impression.
> Or you could take points that are some random number between 1 and 2000 
> apart.
> Do it several times.
> If the plot looks the same each time, you have some info to go on.
>
> It cannot be that each of the millions of points has something to 
> contribute, for the
> reason I gave previously.

Thanks for your hint, but I have some better algorithms.  Perhaps
taking its second order differenial and scan through that, but
it's not easy to implement a perfect one.

Yes for now I just want to have an impression, but as I go
further I may need to analyze them more carefully, better
algorithms may be needed.

> I have frequently found that persons approach solving a problem by computer
> with insufficient consideration for the limitations of computing, 
> requiring huge
> amounts of work for negligible payoff, or involving long sequences of 
> arithmetic
> without considering that the result may have lost all significant digits 
> to essentially
> random bits of round-off.
>
> Sometimes this doesn't matter because the computation is supposed to mirror
> some physical phenomenon which cannot ever be exhibited in reality and whose
> further exploration by computer simulation is essentially irrelevant 
> except as an
> academic exercise.  Computer algebra systems can sometimes be used for
> such pointless exercises where usual numerical computation can't be used.
> You haven't told us anything much about your exercise and whether there
> is some reason you are using Maxima instead of (say) Matlab.

Yeah you guys are assuming I'm doing an excersise.  But it's,
rather, part of my research.

The point is that I want to use Common Lisp, and Maxima seems
doing it natively, I want to give it a try.  I know those FFI
interfaces to other numerical systems, and if Maxima turned me
down, I'd go for them.

> The fact that something is not in a system (or is not documented...) or
> "doesn't seem to work" is occasionally a useful hint that maybe it is
> not a sensible thing to do.

So numerical analysis is not important in a computer algebra
system, right?

-- 
Carl Lei (XeCycle)
Department of Physics, Shanghai Jiao Tong University
OpenPGP public key: 7795E591
Fingerprint: 1FB6 7F1F D45D F681 C845 27F7 8D71 8EC4 7795 E591
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 489 bytes
Desc: not available
URL: <http://www.math.utexas.edu/pipermail/maxima/attachments/20120514/9cf9ffb2/attachment.pgp>;