Posts: 1,665
Threads: 142
Joined: Jan 2009
Using CAS in programs is tricky. I need to store a value inside a CAS variable inside a program using variables local to the program or passed as parameters to the program, I've tried using the CAS command to no avail, since it evaluates to zero.
Assuming aaaa is a CAS variable and A,B,C are variables than have been assigned nonzero numeric values CAS("aaaa:=A*B*C") stores zero in aaaa, but CAS("aaaa:=2") stores the number 2. Also, if I store 2 in aaaa directly in CAS, it successfully retains this value in the program. So, basically, there seems to be no way to store computed values in a program in a CAS variable.
Posts: 709
Threads: 104
Joined: Nov 2005
Quote:
Using CAS in programs is tricky. I need to store a value inside a CAS variable inside a program using variables local to the program or passed as parameters to the program, I've tried using the CAS command to no avail, since it evaluates to zero.
Assuming aaaa is a CAS variable and A,B,C are variables than have been assigned nonzero numeric values CAS("aaaa:=A*B*C") stores zero in aaaa, but CAS("aaaa:=2") stores the number 2. Also, if I store 2 in aaaa directly in CAS, it successfully retains this value in the program. So, basically, there seems to be no way to store computed values in a program in a CAS variable.
You keep forgetting that the CAS currently does not recognize local variables. So CAS("aaaa:=A*B*C") actually creates a GLOBAL variable called aaaa and this GLOBAL variable will have the correct value. However, since local variables have higher priority, the LOCAL variable aaaa still has a value of 0 and that is what is visible to the program. So to do what you want, use:
aaaa:=CAS("A*B*C");
which takes the result of the CAS calculation A*B*C and saves it into the LOCAL variable aaaa.
Here's that old thread again :)
Edited: 12 Nov 2013, 8:42 p.m.
Posts: 1,665
Threads: 142
Joined: Jan 2009
Thanks, but that doesn't solve my problem. It just creates a nul value in aaaa. I need this to be a CAS value not a local value.
Oh dear, now my calculator has crashed again, and I've lost my CAS history, variables, settings... What a pain this thing is to do anything.
Edited: 12 Nov 2013, 9:26 p.m.
Posts: 709
Threads: 104
Joined: Nov 2005
Quote:
Thanks, but that doesn't solve my problem. It just creates a nul value in aaaa. I need this to be a CAS value not a local value.
Oh dear, now my calculator has crashed again, and I've lost my CAS history, variables, settings... What a pain this thing is to do anything.
Are the variables aaaa, A, B,and C all global variables? or are they all local variables? or a mix? This makes a big difference.
If you are using A, B, and C as local variables, then that would not work in CAS. In general, I try to avoid using the builtin variables as local variables. Are you using EXPORT on aaaa or did you leave that out?
I am able to create a CAS variable just fine with
EXPORT TEST()
BEGIN
CAS("aaaa:=A*B*C");
END;
When I go into the memory browser, I see aaaa listed as a real value under the CAS variables. Its value is nonzero, provided that my A, B, and C builtin variables are nonzero. Perhaps there are details I am not understanding about your code.
Edited: 12 Nov 2013, 10:02 p.m.
Posts: 1,665
Threads: 142
Joined: Jan 2009
I'm afraid my calculator is broken again and may need another reformat, so I may just throw it out with tomorrow's trash pickup and cut my losses.
Posts: 193
Threads: 10
Joined: Mar 2008
Hello,
export test(A,B,C)
begin
CAS.abcd:=A+B+C;
end;
seems to work. But I am not sure if it will work on your earlier revision. Can you please try it?
Else, you might want to try
export test(A,B,C)
begin
D:= A*B*C; // note the use of a GLOBAL temporary variable. you could use an exported variable also
CAS("abcd:=D");
end;
for your locking machine. did you try rebooting with backspace pressed? it might be enough to restart it. then erase the offending files...
Cyrille
Posts: 1,665
Threads: 142
Joined: Jan 2009
Well, I've finally gotten everything to work, but it hasn't been easy. As you can see, my dilemma was getting CAS and nonCAS stuff to work together harmoniously inside a program. I'm creating and solving a symbolic expression, which contains coefficients that are evaluated at runtime to be used in the construction of the expression. All of this would be so much easier if there were some sort of CAS switch that could be enabled inside a program at runtime, i.e. CASmode(On  Off). Anyways, here's the code I've written; note the complete absence of LOCAL variables:
Program Beam
//Eigenvalue solver for root based on guess "G"
//The guess should be a positive number slightly larger than the lowest nontrivial (zero)
//root. If it is too low, the result will be a very small number representing zero. Once the lowest
//root is determined, the program can be run again w/o entering the beam data again to
//determine additional higher roots if needed, which will be stored in the matrix variable M0.
//Note that the natural frequencies are proportional to the square of the eigenvalues.
EXPORT Beam_Solve(guess)
BEGIN
G:=guess;
CAS("M0:=nSolve(DET(m1)=0,x=G)");
RETURN M0;
END;
//This program solves for the natural frequencies of a beam with flexible end supports
EXPORT Beam_Main(k1,k2,k3,k4,EI,m,L,guess)
BEGIN
//Dimensionless coefficients
A:=k1*L^3/EI;
B:=k2*L/EI;
C:=k3*L^3/EI;
D:=k4*L/EI;
//Create CAS matrix elements
CAS("a11_:=x^3");
CAS("a12_:=A");
CAS("a13_:=x^3");
CAS("a14_:=A");
CAS("a21_:=B");
CAS("a22_:=x");
CAS("a23_:=B");
CAS("a24_:=x");
CAS("a31_:=x^3*COS(x)C*SIN(x)");
CAS("a32_:=x^3*SIN(x)C*COS(x)");
CAS("a33_:=x^3*COSH(x)C*SINH(x)");
CAS("a34_:=x^3*SINH(x)C*COSH(x)");
CAS("a41_:=x*SIN(x)+D*COS(x)");
CAS("a42_:=x*COS(x)D*SIN(x)");
CAS("a43_:=x*SINH(x)+D*COSH(x)");
CAS("a44_:=x*COSH(x)+D*SINH(x)");
//Create CAS eigenvalue matrix
CAS("m1:=[[a11_,a12_,a13_,a14_],[a21_,a22_,a23_,a24_],[a31_,a32_,a33_,a34_],[a41_,a42_,a43_,a44_]]");
//Solve for eigenvalue based on guess
Beam_Solve(guess);
//Calculate natural frequency in Hz
L0:=sqrt(EI/m)*(M0/L)^2/(2*pi);
//Print eigenvalue and natural frequency
RETURN {M0,L0};
END;
Edited: 13 Nov 2013, 3:51 p.m.
