"Monoids" is outdated

[Pi-Cla]

In: Functors, Applicative Functors and Monoids it described Monoids as having a minimal definition of mempty and mappend. Nowadays one has to define the Semigroup first and then define just mempty to create a Monoid. So people should not manually define mappend anymore. Perhaps a section on Semigroups can be created out of the existing sectiom for Monoids?

[Pi-Cla]

Also, we should in general replace any mention of mappend with <> since the book mostly uses mappend in infix form anyways and this note in the Prelude docs about GHC supposedly removing mappend in some future release.