data-default WAS: [containers] Proposal: Change to the Data.Map Monoid

John Wiegley johnw at newartisans.com
Thu May 22 20:04:56 UTC 2014


>>>>> João Cristóvão <jmacristovao at gmail.com> writes:

> The case is that for very long and complex structures, filled with trivial
> values, and with an obvious default value, writing that default value is a
> task best left for the compiler, with the help of generics.

Ah, I see your point.

Perhaps a compromise solution is possible: A given type can have a generically
defined Trivial value if all its members are either also Trivial, OR they have
Monoid instances with an mempty value.

This way, I know that any Trivial value tree is, at its leaves, composed of
trivial or mempty values, giving a principled meaning of 'emptiness', even
though the Trivial instance may not itself form a Monoid.

The Trivial class would further not allow explicit instances.  The core types
like Int would have hard-coded trivial values in 'base'; but if you want to
define triviality for an ADT in some specific way, it would require making a
Monoid instance to demonstrate what emptiness means relative to an operation.

John


More information about the Libraries mailing list