@Arka The expression can't be factored other than factoring out the constant 16. You can easily prove that (pp1^2+pp2^2+pp3^2) is not a factor by evaluating a at integer values for all the variables and then checking whether the result is divisible by (pp1^2+pp2^2+pp3^2):

A:= eval(a, [Q1=3, Q2=5, Q3=7, Q4= 11, pp1= 13, pp2= 17, pp3= 19]);
                          53328924928
P:= eval(pp1^2+pp2^2+pp3^2, [pp1= 13, pp2= 17, pp3= 19]);
                              819
A/P;
                          53328924928
                          -----------
                              819    

@Arka Then the expression is not factorable as entered. Since you posted your expression as an image rather than as plaintext, I cannot work with it.

@Markiyan Hirnyk What is your point? The options numpoints and gridrefine are not mutally exclusive; options numpoints and grid are. Option numpoints determines the initial grid.

@Carl Love Continuing from above, arbitrarily select omega = 12, so omega^(-1) = 17. Then the matrix DFT is

DFT:= Matrix(4,4, (i,j)-> 12^(i*j) mod 29);

and the inverse transform matrix is

IDFT:= Matrix(4,4, (i,j)-> 17^(i*j) mod 29);

You can verify that their product is 4I by

DFT.IDFT mod 29;

@abbeykabir Sorry, I can't think of any reasonable way to plot a parametric curve over the complex numbers. If it were possible to express it as y = f(x), then there are some possibilities. But that is not possible for this x and y.

@sunit Ah, I see. Was the restart in the .m files? Then that's the problem. The restart only works at the top level, not in a file, and it should be placed in its own execution group.

@sunit So, essentially, you were making the assumptions twice.

@Joe Riel Thanks, Joe. I corrected the Answer above.

@Preben Alsholm Preben, Your eval trick does not work for me. I still have kernel death. I also have kernel death if I replace add with seq or mul, or if I make the equivalent construction with $. Indeed, if I make the line the simple 

C[i]*d[i] + C[i];

then I still have kernel death.

@Preben Alsholm You asked:

By T(N^2(z)) do you mean T(N(z)^2) or do you mean T(N(z))^2?

Neither. By T(N^2(z)) the OP means T((N@@2)(z)), and by (T(z))^2 the OP means T(z)^2.

@Alejandro Jakubi He is saying that the procedure BlahMethod, the export of the module, is not (explicitly) executed. Nonetheless, that is the procedure that contains the statement that causes the crash.

@abbeykabir Yes, I got the same plot.

@Thomas Richard Thanks, I was aware of odetest. But there is a simplicity and obviousness to plugging a solution back into the original that appeals to me.

@Alejandro Jakubi No, I haven't seen any adequate documentation of the stop button. However, I've noticed three things that make it inadequate (I am nearly sure about these, but I haven't seen any documentation).

1. The button will only stop a process during a garbage collection. Ordinarily there is a garbage collection every second or two, so this isn't ordinarily a problem. But if code is very memory efficient, it can be several minutes before there is one.

2. The button will only stop the kernel, not the GUI. This is very frustrating when you've accidentally entered a quick command that produces thousands of lines of output.

3. The button's effect is unpredictable and usually undesirable when applied to multithreaded code.

@nm It is a problem. It is often complained about.

I indeed did retype the OP's expression, although I wish that I didn't need to.