Am 02.04.2013 01:55, schrieb :
> Generally there is a preference for defining clearly things rather than
> giving examples, where at all possible. Specifying something and then
> giving examples is a practice that has the potential for introducing
> additional ambiguities. You can also run into situations where the
> examples and the specification contradict each other.
Appeared to me that the removal of the examples in draft stage caused
semantic loss because a natural language lacks precision. Adding
precision to language adds new ambiguities while a single example could
illustrate expected interpretation. No need to be as more precise as
long as the examples were there.
Leaves the task to verify carefully a semantic loss in the spec, case of
the sort, for instance an example with a negative value demonstrates
that a negative value is permitted, but you don't always find that
specified.
--- A
