Detailed comments on math:e, sinh(), cosh(), tanh()

[michaelhkay]

I should have made these comments before we accepted the proposal, but it's only minor details.

In the example given for math:e, the explanation of the example as a compound interest calculation seems a bit simplistic. There are all sorts of assumptions here about the initial investment, the frequency at which interest is calculated, etc. It might be better just to give the expression and the result and not attempt an interpretation.

According to IEEE 754-2008 table 9.1, sinh() can produce overflow or underflow, cosh() can produce overflow, and tanh() can produce underflow. We seem to be catering for exceptions that cannot occur?

The example results should perhaps be tagged as approximate to ensure that they pass automated testing.

We should perhaps be referencing IEEE 754-2019 (though I'm reluctant to purchase a copy...)