Hey Joe, all, Got it. Thanks!
An associated question: In programming a local "helper" or "auxilliary" function such as dec2bin' in dec2bin :: Integer -> [Integer] dec2bin n = dec2bin' n [] where dec2bin' n acc | n == 0 = acc | otherwise = let r = rem n 2 m = div (n - r) 2 in dec2bin' m (r : acc) is there any way to assign a type signature to the helper function? Michael --- On Sun, 6/28/09, Joe Fredette <jfred...@gmail.com> wrote: From: Joe Fredette <jfred...@gmail.com> Subject: Re: [Haskell-cafe] What is an "expected type" ... To: "michael rice" <nowg...@yahoo.com> Cc: "Haskell Cafe mailing list" <haskell-cafe@haskell.org>, beginn...@haskell.org Date: Sunday, June 28, 2009, 11:29 AM When Haskell runs it's type checker, it tries to "guess" the type of each function. Thats why you can write: map (+1) and it knows that you're talking about a function of type: Num a => [a] -> [a] Another thing, called 'defaulting' resolves this, but you didn't ask about that, so I won't go into it. An expected type is one that you provide to the compiler in the form of a type signature, this can be used to specialize a general type (like the one I showed) or to resolve ambiguous types the compiler can't, or just for documentation/good practice. So when I write: foo :: Num a => [a] -> [a] foo ls = map (+1) ls The "expected type" for `foo` is `Num a => [a] -> [a]`. I imagine you're asking this because you got an error which said your expected type doesn't match your inferred type. That might, for instance, happen if I wrote: bar :: String bar = 'a' 'a' has type `Char`, since `String` is not `Char`, the type checker infers that 'a' has type char, but _expects_ it to be type String. Two solutions are as follows: --- Method 1 bar :: Char bar = 'a' --- Method 2 bar :: String bar = "a" Can you see why those two changes fix the problem? Also, just as a matter of process, I forwarded this to the haskell-beginners list, as I imagine type errors like these come up a lot, and someone probably has a better explanation over there. /Joe michael rice wrote: > as opposed to an "inferred type"? > > Michael > > > ------------------------------------------------------------------------ > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe@haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
