Hi Brett, To the best of my knowledge the interface and implementation of Fran and FRP were not organized around monads. There was briefly an experimental "New Fran" used by FranTk that used monads for implementation. That implementation had some fundamental issues with referential transparency so was largely abandoned. Nevertheless, Yampa does reflect a rather different approach to embedding a dataflow DSL in Haskell than earlier Fran and FRP designs. I wrote an informal account of some of the issues with Fran and FRP that motivated our move to an Arrows-based interface in Chapter 3 of my thesis: http://www.apocalypse.org/~antony/work/pubs/ac-thesis.pdf Hope that helps. Best Regards, -Antony On 2/3/06, Brett G. B. Wortzman <wortzman@fas.harvard.edu> wrote:

Hi all,

I'm doing my undergraduate thesis on monads, comonads, and arrows and their applications. I've been focusing part of my research on the Yale Haskell group's work on Functional reactive programming, and specifically on their move from a monadic to an arrow-based implemenation (as in http://haskell.cs.yale.edu/yale/papers/haskellworkshop02/index.html). I've been unable to find any literature that really gives a good explanation of why they chose to make this move, however. Can anybody shed some light on why arrows were seen as a better choice for this?

Also, does anyone know if there's been any research about implementing FRP with comonads? I can't find any literature about it, but I'm not sure if it's never been looked into, or just attempted and determined to be a bad idea. If the latter, I'd be interested to know why.

Thanks, Brett Wortzman _______________________________________________ Haskell mailing list Haskell@haskell.org http://www.haskell.org/mailman/listinfo/haskell