10/18/2007, 04:11 AM
I've been chastized before for not stating certain things I took to be implicit.
1. This code is only for base e, though it can be modified for other bases (and I'll do so eventually).
2. The solver only gives you the coefficients for the non-constant terms. Therefore, the first element of the vector (e.g., coeffs_100_accel[0]) is the coefficient for z^1. You'll need to insert your own coefficient for the constant term, typically a value of -1. I don't usually need the constant term, because when I've graphed the slog, I've used Runge-Kutta to integrate the first derivative of the slog, which would eliminate the constant term anyway. (This code will be in a later version of the library, by the way.)
3. There are typos in the comments, e.g., the usage instructions that can spit out of Prepare_Vec2 (I added the afix argument later when I converted it from a global variable to a parameter). There are probably typos in the "instructions" as well. Consider this a draft, and be gentle when pointing out inconsistencies and/or errors.
1. This code is only for base e, though it can be modified for other bases (and I'll do so eventually).
2. The solver only gives you the coefficients for the non-constant terms. Therefore, the first element of the vector (e.g., coeffs_100_accel[0]) is the coefficient for z^1. You'll need to insert your own coefficient for the constant term, typically a value of -1. I don't usually need the constant term, because when I've graphed the slog, I've used Runge-Kutta to integrate the first derivative of the slog, which would eliminate the constant term anyway. (This code will be in a later version of the library, by the way.)
3. There are typos in the comments, e.g., the usage instructions that can spit out of Prepare_Vec2 (I added the afix argument later when I converted it from a global variable to a parameter). There are probably typos in the "instructions" as well. Consider this a draft, and be gentle when pointing out inconsistencies and/or errors.
~ Jay Daniel Fox