Ok, that is fine and what I expect as well without manual setting g=h^3
and fiddling around.

But it works not quite yet in cases which are like rational, like some
(positive) constant instead of the '3':

  convert(f, string): StringTools[SubstituteAll](%,"3","Pi"): parse(%):
  simp(%);

Is it that the group laws x^a*x^b=x^(a+b) for the exponent are not used?
Because x^a*y^a = (x*y) ^a is problematic and not always true (assuming
that we talk about (complex) numbers anyway)? Or (x^r)^s?

Aha, I was not aware of that (and hopefully will not forget all these special cases ...)

arg(0) = 0 as convention, so for w := z / abs(z) we have arg(w) = arg(z), 
since only the angle counts (Maple agrees)

Now arg(w) = +-Pi <==> w = -1, so arg(z) = +-Pi <==> z < 0

Which gives the condition for arg(-z) and arg(1-z) and combining interval (and complements)
on paper is easy.

But Maple does not do readable simplifications for intersectiing the ranges from the inital
post. What a pity.

Aha, I was not aware of that (and hopefully will not forget all these special cases ...)

arg(0) = 0 as convention, so for w := z / abs(z) we have arg(w) = arg(z), 
since only the angle counts (Maple agrees)

Now arg(w) = +-Pi <==> w = -1, so arg(z) = +-Pi <==> z < 0

Which gives the condition for arg(-z) and arg(1-z) and combining interval (and complements)
on paper is easy.

But Maple does not do readable simplifications for intersectiing the ranges from the inital
post. What a pity.

Thank you, I was looking for something like 'convert' between the notations
in polar or cartesian coordinates + simplifications for ranges.

It seems it is better to use paper + pencil.

Thank you, I was looking for something like 'convert' between the notations
in polar or cartesian coordinates + simplifications for ranges.

It seems it is better to use paper + pencil.

In the directory of the given link there are some more examples (which I did in a yahoo group, which meanwhile run dry unfortunately), when I tried to learn and understand some of that coding. Also if you search for "LCC" you can find 2 hits, where I applied that.

But I am really curious and looking forward to your post, as using that for Compiler:-Compile is certainly much more interesting than playing with some personal DLLs :-)

In the directory of the given link there are some more examples (which I did in a yahoo group, which meanwhile run dry unfortunately), when I tried to learn and understand some of that coding. Also if you search for "LCC" you can find 2 hits, where I applied that.

But I am really curious and looking forward to your post, as using that for Compiler:-Compile is certainly much more interesting than playing with some personal DLLs :-)

I do not all understand what you want to say, sorry (BTW: what is your native language?), but my English is not very sound, it is more Germish :-)

However: why do you do not need a 'wrapper', have you tried without?

May be you can write down the function (in Fortran, sigh ..) in the sense, that one can see what the needed inputs are. It is 'only' important, whether you update the variables within your code or not. If you are updating you should not use a usual variable, but an array (with 1 element).

But I am not a good programmer ...

What I have seen be looking around: callbacks are possible in Fortran as well

I do not all understand what you want to say, sorry (BTW: what is your native language?), but my English is not very sound, it is more Germish :-)

However: why do you do not need a 'wrapper', have you tried without?

May be you can write down the function (in Fortran, sigh ..) in the sense, that one can see what the needed inputs are. It is 'only' important, whether you update the variables within your code or not. If you are updating you should not use a usual variable, but an array (with 1 element).

But I am not a good programmer ...

What I have seen be looking around: callbacks are possible in Fortran as well

Thx :-)

I think that holes should possible by using piecewise( ..., NULL) or piecewise( ..., " ") instead of max, dito min.

Thx :-)

I think that holes should possible by using piecewise( ..., NULL) or piecewise( ..., " ") instead of max, dito min.

Ah, ok ... having WIN XP SP2 with Maple 12 I never had such instability
and use a modified version of "Classic Windows" Windows theme.

The only issue was connecting my Flat screen, where 'digital' gave troubles
after a release change (but it was impossible to reconstruct wether some
driver has been changed).

What occassionally happened for older sheets: I was not able to open it
in newer versions: I first had to remove all outputs, save it and then
it works.

Ah, ok ... having WIN XP SP2 with Maple 12 I never had such instability
and use a modified version of "Classic Windows" Windows theme.

The only issue was connecting my Flat screen, where 'digital' gave troubles
after a release change (but it was impossible to reconstruct wether some
driver has been changed).

What occassionally happened for older sheets: I was not able to open it
in newer versions: I first had to remove all outputs, save it and then
it works.

The only 'pattern' I see is:

He takes a polynomial of degree 3 with rational coefficients in n. It has 1 real root function f in n (test at n=1).
Now he is summing f(n) over n

The only 'pattern' I see is:

He takes a polynomial of degree 3 with rational coefficients in n. It has 1 real root function f in n (test at n=1).
Now he is summing f(n) over n