Thank you, Armando and Jer, for your very clear explanations, in particular explaining why there is such a problem with 42*2^(1/2).
Interestingly, working from the opposite direction, WolframAlpha doesn't usually offer large multiples of square roots as a possible closed form, but 42*2^(1/2) comes up at once as a possibility for 59.3969696.
So I guess it is a 'known issue' in those circles, at least.
Posted by broll
on 2016-06-11 20:50:40