> If it makes sense to apply a "Foo of functions" to a value,

That's just a functor. "fmap ($ value) foos".

But yes, if you want to do this with more than one foo (or functions inside a foo), then you need applicative.

e.g.

(+) <$> [1,2,3] <*> [10,20,30]

On Wed, Jun 15, 2016 at 10:29 AM, John Wiegley <johnw@newartisans.com> wrote:

>>>>> martin <martin.drautzburg@web.de> writes:

> I am at a stage, where I can use some of the Monads and Applicatives which
> are out there. But I hardly ever wrote my own instances. I am curious to
> learn about the thought processes which lead to the insight "hey that can be
> written nicely as an Applicative Functor"

When I realize Monad might be useful:

If it makes sense for a "Foo of Foos" (over some type) to be reduced to a
Foo.

When I realize Applicative might be useful:

If it makes sense to apply a "Foo of functions" to a value, or a function
taking two or more arguments to two or more Foos, throughout those Foos.

--
John Wiegley GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com 60E1 46C4 BD1A 7AC1 4BA2
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe