@Christopher2222 

http://www.maplesoft.com/applications/view.aspx?SID=4049

@Christopher2222 

Though on my brwoser it looks a bit different all the links work (downloading the worksheet  4 MB and preview as hmtl)

@Christopher2222 Fine (may be clearing the browser cache helps?)

Twitter: I meant Maple and corrected my post.

@Christopher2222 Usually I have a quite restricted FF as browser (resulting in difficulties on Maple's pages). But the link works. Without the parameter "view=html" it shows the usual overview for the application, i.e. the links for preview and downloads.

PS: the new Primes appearance is a real mess for posting - it only half the screen width only, shows some lines only. It is like using a Smarty :-(

Why dont you Maple force us to use Twitter to work here?

Agreed, such limitations of 'maximize' shall be documented.

I think that Maple can not find the integral (without help) and maximize does not work on integrals.

The other point is: maximize seems not to work on (certain) specials functions.

@tomleslie it is the indicator function of the unit circle (as 'inner side' of circle line)

Heaviside(1-x^2-y^2);
plot3d(%,   x=-1..1, y=-1..1, axes=boxed);

In this case the option method=ftoc lets it work.

For the (area under the) parabola FTOC has an error (which could be corrected manually, I think).

Look up the series command,

https://en.wikipedia.org/wiki/Taylor_series

I would simplify call this approach "fitting". If one allows a whole "zoo" through combinations then it may be worthless overall. Even if one has some parametric model it is non trivial - but may be valuable.

One may search for "density estimation" in Finance, years ago I 'played' with that. Likewise any field of your taste.

It starts with cleaning input: what to accept or receject or 'correct' and how. The next step is called 'data binning' ... as one hopefully does not have only some dozends inputs. And data may have a time dependency (or similar, may be hidden)

Only after that it makes sense to start at all.

Now you may have some 'nice' fits - but they may not meet what is expected/wanted/experienced in the tails - which may not be directly visible from data. Or other reasons.

It is not even clear that for some model one arrives at unique solutions or those are 'stable'.

And do not forget that such "experiments" might have the goal to 'test' or falsify the model.

Hm ... I do not miss such a feature.

@Christopher2222 changing the date again will not work, there are still "items in the future": sooner or later you have to switch to actual time to have other programs at hand

Searching depends on your machine, operating system and setting. Roughly.

Under Windows you open the file explorer. You can enter a search term (depending on your settings). You may search for *.* , which means for all files or directories as far as they are not hidden.

Sort by date. You will see the directories first. Check the date and make a screenshot (for the names). Then scroll beyond the directories, you will find the files. Do the same.

Tell support about the results and what to do with them.

PS: of course it is a mess how Maple wants to protect its copyrights / license in such an unfirm way.

may be  "%" are meant

@tomleslie  My default is Digits = 15 and with that it does not work. I have to set Digits:=10 after the restart command.

The problem are the poles. A solver may not 'jump' properly to the next regime.

PS: a manual workaround is to consider the numerator only, but checking against the denominator usually has to check against cancelling zeros as well. Or to use fdiscont and work inbetween.

PPS: may be you want to "play" with the more simple re-formulation sin(Pi*t*(2*t-1))/cos(Pi*t)/cos(Pi*t^2)^2 = 0

NB: 'solve' seems to ignore the denominator has to be regarded

@tomleslie for Digits = 15 it runs into a numerical loop, for Digits=10 it works

And for "RootFinding:-Analytic": the function is not very analytic :-)

Still you would have

arctan(a*t^2, t^2):
%=evalc(%) assuming a::real;
                               2   2
                     arctan(a t , t ) = arctan(a)

and that is something you have in your case (in the small), I think