I'm using Set for algorithms which work this way: while the set is non-empty choose an element from the set do something with that element remove the element and some more that depend on it However when I code this with Set.null and Set.findMin, I run into the same trouble like with List.null and List.head, namely Set.findMin is a partial function. I like to have a function like choose :: Set a -> Maybe a choose set = toMaybe (not (Set.null set)) (Set.findMin set) toMaybe :: Bool -> a -> Maybe a toMaybe False _ = Nothing toMaybe True x = Just x in Data.Set. Opinions?

On 12/11/06, Henning Thielemann wrote:

Opinions?

Well, you can do this in your code, no? It seems kind of silly to wrap it up in a Maybe when the resulting code will case on it. Calling code frequently has: if Set.null set then ... else ... (Set.findMin set) ... -- Taral "You can't prove anything." -- Gödel's Incompetence Theorem

You can try Data.Set.minView. http://www.haskell.org/ghc/docs/latest/html/libraries/base/Data-Set.html Cheers, JP. On 12/11/06, Henning Thielemann wrote:

I'm using Set for algorithms which work this way: while the set is non-empty choose an element from the set do something with that element remove the element and some more that depend on it

However when I code this with Set.null and Set.findMin, I run into the same trouble like with List.null and List.head, namely Set.findMin is a partial function. I like to have a function like choose :: Set a -> Maybe a choose set = toMaybe (not (Set.null set)) (Set.findMin set)

toMaybe :: Bool -> a -> Maybe a toMaybe False _ = Nothing toMaybe True x = Just x

in Data.Set.

Opinions? _______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

On 12/11/06, Henning Thielemann wrote:

On Mon, 11 Dec 2006, Jean-Philippe Bernardy wrote:

...

You can try Data.Set.minView. http://www.haskell.org/ghc/docs/latest/html/libraries/base/Data-Set.html

Seems to be new in GHC-6.6 over GHC-6.4.

True, though "base 2.0" is more accurate, in light of the recent fuss on the subject.

I suspect that not really monadic behaviour is needed in minView, but only the fail method, isn't it? Confusing.

Indeed. This is standard style in many modules however. (And an argument for the resurrection of MonadZero.) Cheers, JP.