A possible re-interpretation of the mechanisms of oximation of carbonyl compounds