On Sun, 2008-01-06 at 23:44 +0100, Peter Verswyvelen wrote:
If I understand it correctly, implicit parameters in Haskell allow you to pass values to functions with explicitly adding a parameter to each of the functions being “called” (I appologize for my imperative terminology here. How would I say this correctly? Being “evaluated”?)
No one is going to get upset if you use "call", but if you'd like, the more technical term in this context is "applied", instead of "call f passing x", we would say "apply f to x".
The arrows always use tuples to group the input and output parameters, like:
foo :: SF (Int,Int,Int) (Int,Int)
foo = proc (x,y,z) -> do
p <- cat -< (x,y)
q <- dog -< z
returnA -< (p,q)
where
cat = proc (x,y) -> returnA -< (x+y)
dog = proc z -> returnA -< 10*z
Suppose I don’t want to explicitly pass the x and y parameters to the cat (and deeper) arrows, but make them implicit. I guess that would be impossible? I mean, I can’t use implicit parameters language extension to make the arrow input parameters implicit?
Implicit parameters add an extra argument to a function conceptually. What you need is to "add an argument" to "SF" which implicit parameters don't know how to do since SF is just some data structure. One way to deal with this is the way you deal with the same problem in Haskell without implicit parameters. (I never use implicit parameters). In that case you would use the Reader monad (transformer). Similarly, you can use an equivalent Reader/Environment arrow transformer.