For me as well? Or do I need to have purchased MapleSim?

Is that really the way how Math is learned to day? I would find that frustrating, it like in School, mechanical - but having a CAS at hand (sorry, but do know much about the 'system' in foreign countries) ...

.

Though I understand your desire (more or less) there is no "standard"
for transforms or special functions, not even for Fourier transform.

And not sure whether someone has 'Bracewell (p335)' it would be best
to write down his definition for the transform, dito for the Bessel
function he uses (it is quite often, that different notations and
factors for scaling or norming are used by different authors).

If you have them both and compare that to Maple's convention then you
can simply translate as shown by Alex.

export as html gives me good pictures as *.gifs (working with the Reasonable Interface), it may depend on MS Word how they are imported

This way you divide numerical by zero, exp( verysmall*x ) - 1 ~ 0.


  params:=[c=2.998*10^8,k=1.381*10^(23),h=6.626*10^(-34),t=4500];

  (h/k/t*x); eval(%, params);
  eval(%, x=1)..eval(%, x=100);;

                              -59                       -57
           0.10662161074906 10    .. 0.10662161074906 10

Now xi = h/k/t*x is very small, so you can use a Taylor series,
even with Digits:=14:

  # version 1
  exp(xi)-1; series(%, xi=0,3);
  convert(%,polynom);
  E1:=unapply(%,xi);

                                              2
                       E1 := xi -> xi + 1/2 xi

  (8*Pi*h/(c^3))*(x^3/E1(h*x/(k*t)));
  P1:=eval(%, params);
  plot(P1, x=1 ..100);


  # version 2
  1/(exp(xi)-1); series(%, xi=0,6);
  convert(%,polynom);
  E2:=unapply(%,xi);

                          1                             3
             E2 := xi -> ---- - 1/2 + 1/12 xi - 1/720 xi
                          xi

  (8*Pi*h/(c^3))*(x^3*E2(h*x/(k*t)));
  P2:=eval(%, params);
  plot(P2, x=1 ..100);

For an experienced coder it will be simple, if your specifications are correct (and that is, what he has to work out with you)

perhaps none: MS Access depends on versions (and may be has passwords). The main point however: it stands not for a table, but for a relational database. So may ask the developpers to write a makro for exporting data to a flat file, to be read from Maple. For triggering that you may look up the MS manual to start / stop Access with a flag by a command line command to be sent by Maple.

.

Laurie Anderson (not sure about spelling) had a song and said "there is no Captian"

I suggest you for the most lame homework cry of the last quaters and: "there will be no answer"

try the most simple case: a line in the plane.

- how many points do you need to determine it
- what can you say if you have only 1 point
- how would you solve that instead of your original exercise

then try to start over ... avoid Maple as far as possible for all that
until you got the principle

what have you done so far (by paper and pencil?) to set up equations?

you may want to post your *concrete* example (even if one might guess), it may be something around 'eval' or working with indexed variables

Still I think your way is too brute (and how would you go on if the matrix is populated?).

First use maplemint(A) before assigning the other constants, it will tell you that some variables
are not declared and others are not used (especially mat1i is not used ...)

Then you execute A and create a sequence - but your bound is n3 = 1000000, 1 Million ... and
it is not clear, whether you would ever use the sequence or just the result on the matrix. And
you just call the proc a million times.

Before doing that I would test *small* values (especially a smal size for the matrix), your code
does not allow that easily, you may wish it is possible.

So I called it just once (wanted to look into some coefficients). Already the that gave me an
error: "Error, (in A) unable to store '-1.333333333+.1e-5*ec(8,9)' when datatype=float[8]".

This says that your function ec is not defined. And all you have through this in 300 sec that
your matrix is not properly filled (I am astonished, that you did not receive such a message,
have you ever looked into some parts of your result? When do you think it is correct?).

Concerning memory you may wish to estimate how many entries are non-zero to get some
feeling how many memory would be needed only to store that.

Anyway: my advice would be to start with a small size, check everything is correct there and
then increase for a kind of 'stress test'.

And why 8000000^2 grid points? That would give tremendous runtimes.

Your input is not quite readable and not usable for copy&paste ...

Anyway: without 'sparse' you have 2 matrices of size 8000000^2 * 32/4 * byte ~ 500000 GB and even if they are very sparse you will run into pain by filling them, even if  in a sparse way.

My feeling (which may be wrong, of course) is that your approach is a bit brute and that Maple may be not the right tool for that.