> Precision (readability and repeatability) is something quite
> different from reliability (certainty, belief, trust).
I agree! Nonetheless, I think they both have something in common: they 
seem to have been carefully disregarded by (most of) earlier 

> developing the dataset profiles using our work.
If each application works as standalone, there is no problem, but if we 
focus on interoperability, then it would be really good if each datatype 
means the same thing in every application where it is used.

> What are the points in the continuum?
That is a really GOOD question. I consider that we can not know how many 
points there are within an interval. If we want to express it, we would 
need something more than just something implicit. I consider that we need 
to add a couple of new syntactic constructs to be able to express what we