I'm not sure, but it seems as if workprec in fdiff may have a bug. See an example. Or, maybe, something is wrong with the numerical ode solution there.

Alec


Please Wait...