I tried the NullSpace command with and without the [1..7,1..8]. without it returns {} and with it it returns a massive formula. Like you said 7 vectors I presume each with 8 components ie [1..7,1..8]. Why does one have to give maple the [1..7,1..8]? It doesn't say anything about this in the maple help.

I tried the NullSpace command with and without the [1..7,1..8]. without it returns {} and with it it returns a massive formula. Like you said 7 vectors I presume each with 8 components ie [1..7,1..8]. Why does one have to give maple the [1..7,1..8]? It doesn't say anything about this in the maple help.

thanks, thanks also for text mode tip i had been cutting and pasting my maple expressions from another editor into maple to avoid the gui formatting surprises.

thanks, thanks also for text mode tip i had been cutting and pasting my maple expressions from another editor into maple to avoid the gui formatting surprises.

yes you are right it is entirely symbolic. It is an 8x8 jacobian A(u,w) with the 8 variables,u, and 5 parameters. 3 of which describe a direction w. I have the eigenvalues they are messy but not too bad. The bad part about them is that they have the sqrt in them which I think complicates the work for maple. Although I have replaced the sqrt function with()^(1/2) and specified that all the variables and parameters are real. I am guessing that those assumptions make maples job more difficult. I would like the eigenvectors explicitly I was going to examine the non-linearity of the characteristic fields grad(\lambda(u,w).r(u,w). I am new to maple, and I noticed that removing extraneous parenthesis resulted in a speed up of about 2 for the computation of the eigenvalues, so the thought crossed my mind that there may be other things one can do to ease maples job... for this system the 6 of the 8 eigenvalues come in pairs that differ by the sign of a term, so that if I can find one from each pair I could probably deduce the remaining ones.

yes you are right it is entirely symbolic. It is an 8x8 jacobian A(u,w) with the 8 variables,u, and 5 parameters. 3 of which describe a direction w. I have the eigenvalues they are messy but not too bad. The bad part about them is that they have the sqrt in them which I think complicates the work for maple. Although I have replaced the sqrt function with()^(1/2) and specified that all the variables and parameters are real. I am guessing that those assumptions make maples job more difficult. I would like the eigenvectors explicitly I was going to examine the non-linearity of the characteristic fields grad(\lambda(u,w).r(u,w). I am new to maple, and I noticed that removing extraneous parenthesis resulted in a speed up of about 2 for the computation of the eigenvalues, so the thought crossed my mind that there may be other things one can do to ease maples job... for this system the 6 of the 8 eigenvalues come in pairs that differ by the sign of a term, so that if I can find one from each pair I could probably deduce the remaining ones.

thanks, the effect I am trying to achieve is that maple assumes that every variable(there are 12 in my 8x8 system) is real. Is there an efficient way to tell maple to assume all variables are real?? after your advise I now have to LinearSolve(A, b,free='S') assuming real, S::real; and maple makes a big mess, inserting a bunch of assume statements into the answer. But at least I can be sure maple isn't giving me some complex functions back, thanks again.

thanks, the effect I am trying to achieve is that maple assumes that every variable(there are 12 in my 8x8 system) is real. Is there an efficient way to tell maple to assume all variables are real?? after your advise I now have to LinearSolve(A, b,free='S') assuming real, S::real; and maple makes a big mess, inserting a bunch of assume statements into the answer. But at least I can be sure maple isn't giving me some complex functions back, thanks again.

no affect

no affect

other forum had this to say post I understand where he is coming from I think he's the issue pinned down.

if you plug maple's suggested solution:c=0 into the original equation then the equation simplifies down to: 0 = a^2+b^2 which is not by itself a solution! it is only a solution if a = ib or b = ia. So maples idea of a solution is actually another equation... like wise if you plug maples other suggested solution in to the original eqn you end up with: 0 = (a^2+b^2)*a^2/(a^2+1) which is only a solution if a!=i, and a=ib or a=0 I was expecting the result that maple produced to reduce the rhs of eqn to 0...