[Haskell-cafe] Do you recognize this type?

satvik chauhan mystic.satvik at gmail.com
Fri Jun 20 15:47:19 UTC 2014


I might be wrong but what is the problem in folding with the `SumEvent`
over all the events?


On Fri, Jun 20, 2014 at 8:40 PM, Corentin Dupont <corentin.dupont at gmail.com>
wrote:

>
>
>
> On Fri, Jun 20, 2014 at 4:55 PM, Chris Warburton <
> chriswarbo at googlemail.com> wrote:
>
>> Corentin Dupont <corentin.dupont at gmail.com> writes:
>>
>> > Nobody on that?
>> > What I really want is: a computation that can be interrupted if enough
>> > result is already there.
>> > I would be surprise if that concept doesn't exist already somewhere :)
>>
>> I'd be surprised if this can be done with Applicative without some
>> hackery.
>>
>> Very roughly, we can think of Functors as one-shot actions. For example,
>> the IO instance of Functor lets us perform exactly 1 side-effect; all we
>> can do with the result is map pure functions over it.
>>
>> Applicatives are a static chain of actions. For example, the IO instance
>> of Applicative lets us chain together any number of side-effecting
>> actions, but those actions can't alter the chain itself.
>>
>> Monads are a dynamic chain of actions. They're like Applicatives, but
>> each action can choose which action to perform next based on the
>> previous results.
>>
>
> wow, I *love* this summary! Thanks for that!
>
>
>>
>> From this perspective, the "shortcut" you're looking for requires the
>> ability to choose the next action (count another vote or finish
>> prematurely) based on the previous results (the running total). That's
>> what Monads are good for, but Applicatives can't do it by themselves.
>>
>
> Indeed I could make my DSL a monad instance with some changes, that would
> be easy to do. But even not a Monad cannot do the trick: imagine if in a
> country, you were asking
> the people to vote one by one, only to interrupt them when the result is
> known. In fact they have to vote altogether,
> in whatever order, and we count the votes as they arrive.
>
> So I think what I need is a sort of Applicative++: not a chain of actions
> but a set of actions, all independent from each other (so no monad) and
> with no particular order.
>
>
>
>
>>
>> There's probably an unsafe* way to do this, which will in fact be safe
>> due to your voting invariant ("if there aren't enough uncounted votes
>> to change the outcome, the outcome cannot change"). Maybe someone who's
>> performed similar optimisations can help.
>>
>> Good luck!
>>
>> Chris
>>
>
>
> _______________________________________________
> Haskell-Cafe mailing list
> Haskell-Cafe at haskell.org
> http://www.haskell.org/mailman/listinfo/haskell-cafe
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/haskell-cafe/attachments/20140620/bb2f1592/attachment.html>


More information about the Haskell-Cafe mailing list