Well, first, you can use the default ones by using:

Setup(tetradmetric = orthonormal) #or
Setup(tetradmetric = null):
e_[];

that will give you the appropriate tetrads corresponding to your metric signature.

For more complicated metrics, that procedure may fail. But you can still input them as:

l[mu] = [1/2*sqrt(2)*sqrt(r)*I/sqrt(r - 2*m), 0, 0, -I*1/2*sqrt(2)*sqrt(r - 2*m)/sqrt(r)],
n[mu] = [1/2*sqrt(2)*sqrt(r)*I/sqrt(r - 2*m), 0, 0, 1/2*sqrt(2)*sqrt(r - 2*m)*I/sqrt(r)],
m[mu] = [0, -1/2*r*sqrt(2), -I*1/2*r*sin(theta)*sqrt(2), 0],
mb[mu] = [0, -1/2*r*sqrt(2), 1/2*r*sin(theta)*sqrt(2)*I, 0];
Define(%)

(that's for Schwarzshild according to my notes)

You can also redefine eta_[] used to calculate the tetrads trough say:

new_eta:=Matrix(1..4,shape=symmetric,{(1,1)=2,(1, 2) = 4, (3, 3) = 1,(4,4)=a});
Setup(tetradmetric=new_eta)

The example is just a random matrix, but the idea is to reproduce non-holonomic bases.

I have 4GB of Ram and I still couldn't open the file at all.

you know, I was thinking of something like this, but when I checked the maple help, it gave me different use for the while statement. I'll try your way, it might be better then mine. Anyway, I'm glad I have now two solutions instead of zero that was stalling me for the whole past week. Thanks...

Scott, Could you re-post your code what you wrote with > and < (here are the HTML codes-you can find them in the "Input format") for the greater and less signs,cuz this way it's not very understandable for me at least... What I did was to put one independent counter like for m from 1 while (a <1) do a:=a+p if ... end do; Meaning I manipulate the step outside the cycle, and the counter (m) serve for other purposes. It seems to be working. I think it's more or less the same as what I supposed, though it's not very visible :) But the question is why it should be so hard, when it can be simple. If the step is being checked on every repeat and not just memorized in the beginning.