I've changed my procedures by their specifications. I just want to report this trouble: when I think about specifications,
I naturally think in terms of sets, when I think about algorithms, I naturallly think in terms of arrays
(and then in terms of finite sequences) and then when I'm juggling with specification on one hand and algorithms
on the other one, I need to transform the formers into the latters.
In fact sets are not natural in algorithmics.
--
FL