Thanks Ray,
I am a novice when it comes to Maxima. A student of mine asked if he
could see how the sum started to converge into a sequence of impulse
functions, and I thought I would try and use Maxima instead of octave on
a whim, and you know, I like it. I didn't even realize that it would
simplify into such a nice trigonometric form, but I suppose I should
have suspected it, looking at the plot. I guess the blank areas grabbed
my interest instead.
I vote for a larger default value of nticks. Most folks would expect a
constant number of points/length on the x axis, and having a constant
value in the y dimension is not something that they would expect from
that algorithm. Speed is secondary to accuracy, in my opinion, and the
folks who understand how the adaptive algorithm works can put nticks to
1 if they like for speed. I think at least 100 would be a good default
value. The screen is inherently discrete and users know this, so having
something that is on the order of the size of the x axis on the plot
window would be good.
Thanks,
Rob
On Sun, 2007-12-02 at 10:45 -0500, Raymond Toy wrote:
> Rob Frohne wrote:
> > Hi Everyone,
> >
> > I find a funny thing happening when I plot with this command:
> >
> > plot2d([realpart(ev(sum(%e^(-2*%i*%pi*f*k),k,-50,50),simpsum))], [f,-2,2],[y,-50,50]);
> >
> >
> Others have answered the question about plotting.
>
> Did you know that nusum can produce a better answer for this summation?
> trigreduce(trigsimp(realpart(nusum(exp(-2*%i*%pi*f*k),k,-50,50))))
> produces csc(%pi*f)*sin(101*%pi*f). This gives a better feeling for
> what the graph ought to look like, is faster to plot, and probably has
> fewer round-off errors.
>
> Ray
>