Re: mathlib

[Roberto Ierusalimschy <roberto@...>]

> But since that is not on the table, the main argument in favour of
> precisely sinh, cosh, tanh is that they belong to a class of functions
> called "elementary" by mathematicians, which are commonly taught
> even to engineering students in English-speaking countries, and
> that their implementation involves subtle points.

That is a good argument. Do you have any reference about that (e.g., a
list of those "elementary" functions)?

-- Roberto