I have dusted off the some first order ode routines I wrote last year and am
trying to get them into shape for submission. I am after opinions on a couple
of design decisions:
1. The exisiting ode2() routine returns a single solution. I am proposing to
return a list of solutions, similar to solve, as non-linear odes can have
multiple solutions.
dy dy 2
(contrib_ode(eqn : - y + 1 - (2 - x) -- + (--) = 0, y, x);
dx
2
2 x - 4 x
[y = %C x + %C - 2 %C + 1, y = - --------]
4
2. Some of the solutions are parametric, and I need to chose the
parametric variable - the example below uses p.
I am considering using %t, and hardcoding the choice. Thoughts?
dy 2 dy
contrib_ode(eqn : (--) x -- - y = 0, y, x));
dx dx
2
3 LOG(p - 1)
- -------------
2 3
[x = %C %E , y = p x]