Better implementation of Show instance for Data.Scientific

[Henning Thielemann]

On Sat, 2 May 2015, Phil Ruffwind wrote:

> OTOH, I think it makes sense to automatically drop the '.0' when the 
> number is integral, and that could be built into the Show instance. Of 
> course, there's the question of whether such a change would break other 
> packages...

I think if a program relies on a certain formatting in the result of 
'show', it is just a bug. Btw. Show should always show valid Haskell 
expressions as "deriving Show"-generated implementations do.