Yitzchak Gale writes:

Tom Davies wrote:

...

...

I find it convenient sometimes to convert a Maybe value to an Either

maybeToEither = flip maybe Right . Left

Remember, some people don't like flip! :p maybeToEither = (`maybe` Right) . Left

While useful, I think its ubiquity to simplicity ratio is not high enough to justify either depending on MissingH just for that, or adding it to a base library.

Just like the swap :: (a,b) -> (b,a) function a lot of people were discussing on libraries@? In general, I agree.

However, if your personal style makes this so ubiquitous as to override its simplicity, go ahead and upload it to Hackage, perhaps others will also benefit.

Just give it a good name, rather than "fooToolkit", "barToolkit", etc. -- Ivan Lazar Miljenovic Ivan.Miljenovic@gmail.com IvanMiljenovic.wordpress.com

Ivan Lazar Miljenovic schrieb:

Yitzchak Gale writes:

...

While useful, I think its ubiquity to simplicity ratio is not high enough to justify either depending on MissingH just for that, or adding it to a base library.

Just like the swap :: (a,b) -> (b,a) function a lot of people were discussing on libraries@?

I have 'swap' in my utility-ht.

In general, I agree.

Problem in Haskell is, that it allows for a high degree of modularization such that most components become trivial. Does that mean that we should not define simple functions at all, which in the extreme case would mean that we inline all potential function calls? I think converting Maybe to Either is an often repeating task which justifies an individual function. Maybe I add it to utility-ht.

On 4 August 2010 18:40, Alexander Dunlap wrote:

It's also nice for people reading code if common functions are functions from common libraries. This allows readers' "vocabulary" of common functions to increase, so they don't have to trawl through someone's personal "utility" library to figure out what each utility function does.

Agreed. That's why I like readMay. There are many ways to write a reads wrapper: case reads str of [(x,_)] -> Just x; _ -> Nothing vs case reads str of [(x,"")] -> Just x; _ -> Nothing -- stricter Etc. It seems everyone defines their one, where as Safe.readMay is common. There could be others defined in Safe that encompass most cases.

On 03/08/2010, at 10:09 PM, Ertugrul Soeylemez wrote:

Tom Davies wrote:

...

I find it convenient sometimes to convert a Maybe value to an Either thus (excuse the syntax, it's CAL, not Haskell):

maybeToEither :: a -> Maybe b -> Either a b; maybeToEither errorValue = maybe (Left errorValue) (\x -> Right x);

As a side note, this is perfectly valid Haskell code, even though it's quite verbose. The equivalent:

maybeToEither :: a -> Maybe b -> Either a b maybeToEither leftValue = maybe (Left leftValue) Right

I like that better, and CAL supports that, thanks.

Honestly I find this better than the completely points-free variants posted in the other subthread. They demonstrate overusing points-free style.

I agree -- as an inexperienced functional programmer, but not a complete beginner, I found the points free versions incomprehensible! Tom

On Tue, Aug 3, 2010 at 8:33 PM, Claude Heiland-Allen wrote:

{-# LANGUAGE TemplateHaskell, ScopedTypeVariables #-} import Language.Haskell.Djinn (djinnD) $(djinnD "maybeToEither" [t|forall a b . a -> Maybe b ->  Either a b|]) main = print . map (maybeToEither "foo") $ [Nothing, Just "bar"]

This is very cool (as is Djinn itself), but for me the ideal syntax would be maybeToEither :: a -> Maybe b -> Either a b maybeToEither = $(derived) Is something like this possible in TH? The splice would have to know its declared (even inferred?) type. --Max