Probability Functions: order of arguments insanity

[parisse]

It's of course true if p and k have values, not if they are symbolic. There is no good solution anyway, since users will expect p as an optionnal 3rd argument to binomial(n,k) and will expect argument order consistency with binomial_cdf/icdf, whatever it means since the third arg has nothing to do with the k of binomial(n,k,p), you should see this k as a disappearing argument.