[Haskell-cafe] FRP, Simulations and Time (Sodium et. al)

David Sorokin david.sorokin at gmail.com
Mon Mar 17 06:01:35 UTC 2014


It depends on that how FRP is defined.

So, in F# it is usually associated with the Async monad and IObservable

If we will follow in the same direction then the short answer is definitely
yes. An example is my simulation library Aivika [1].

My Process monad allows describing the discontinuous processes (inspired by
the Async monad from F#). There is also the Signal type (inspired by the
.NET IObservable interface) that allows notifying about some (.NET-like)
events such as an update of the variable.

Then the characteristic function in a context of FRP would be the next one:

processAwait :: Signal a -> Process a

There is an opposite direction from the Process computation to signals
through the Task type but it is a little bit complicated as the
discontinuous process can be canceled, or an IO exception may arise within
the computation.

But the simulation field is too big and diverse to assert something general.


[1] http://hackage.haskell.org/package/aivika

On Sun, Mar 16, 2014 at 5:23 PM, martin <martin.drautzburg at web.de> wrote:

> FPR is usually described as a way to model interactions with the real
> world. I believe the same ideas should be
> applicable to simulations. Instead of "real" events I would use fake
> events. This however only makes sense when the Time
> associated with the Events (and Behaviors) is not wallclock time, but some
> kind of virtual time.
> I looked briefly into Sodium and found no way to use "my own time". Steven
> Blackheath even said during a presentation,
> that Time in Sodium is just another Behavior. I am not sure if I
> understood this correctly, but it gives me the feeling,
> that the only Time available in Sodium is wallclock time, which would make
> it unsuitable for simulations.
> Could someone give me some insights whether or not FRP is suitable for
> simulations?
> Does my reasoning about Time make any sense? Do different libraries treat
> Time in different ways, such that some are
> suitable for simulations, while others are not?
> Martin
> _______________________________________________
> 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/20140317/73619b4a/attachment.html>

More information about the Haskell-Cafe mailing list