pramp(
    f;
    p = [0, 1],
    hmin = (p[end] - p[begin]) / 1000,
    hmax = (p[end] - p[begin]) / 10,
    h = hmax,
    hgrow = 1.2,
    hdegrow = 0.5,
    verbose = false
)

Run 'f(p)' successively with parameter values from the range given by 'p'. Stepsize is given by h. If f doesn't throw an error, h is increased by the factor of hgrow unless it exceeds hmax, and the next step is performed with the new value.

If f(p) throws an error, solution is retried with a lower value of h=h*hdegrow. In this case, if p==p[begin] or h<hmin, the error is rethrown.