[GHC] #974: Add unzipEithers, lefts, rights to Data.Either

[roconnor at theorem.ca]

On Mon, 13 Nov 2006, Ross Paterson wrote:

> On Sun, Nov 12, 2006 at 11:40:19AM -0500, roconnor at theorem.ca wrote:
>> Everyone seems satified with the name unzipEithers.  There have been
>> suggestions for
>>
>>  partitionEither :: (a -> Either b c) -> [a] -> ([b], [c])
>>
>> but I suggest they be pursued in a different trac.
>
> I think it makes sense to consider them together, as these functions are
> closely related, the only concerns anyone has raised about them relate
> to names, and they're competing for the same part of the namespace.

But is this function (a -> Either b c) -> [a] -> ([b],[c]) defined by 
anyone and/or used anywhere?  If not, is there any evidence that people 
will use this function in the future if it is added?

My reason for adding [Either a b] -> ([a],[b]) is that at least two 
independent projects from independent authors use it and it seems 
generally widely useful.  Althought techinically I don't have the same 
evidence for lefts and rights, I believe such evidence exists.  At least 
one instance is for specifying the behaviour of unzipEithers, which is why 
I include it in the same patch.  However I'm not convinced that there is 
demand fur the (a -> Either b c) -> [a] -> ([b],[c]) function.  This is my 
concern.

Since no one is suggesting the name of partitionEithers for the [Either a 
b] -> ([a],[b]) function anymore, there is currently no name space 
dispute.

For these reasons, I suggest (a -> Either b c) -> [a] -> ([b],[c]) be 
pursued in another trac.

-- 
Russell O'Connor                                      <http://r6.ca/>
``All talk about `theft,''' the general counsel of the American Graphophone
Company wrote, ``is the merest claptrap, for there exists no property in
ideas musical, literary or artistic, except as defined by statute.''