Language extension idea (was Re: [Haskell-cafe] Re: OCaml list sees...)

[Tom Pledger]

oleg at pobox.com wrote:
[...]

>Actually we merely need to add a deconstructor. Also, we can leave the
>type of elements fully polymorphic. Something like this:
>
>>class List l where
>>    nil :: l a
>>    cons :: a -> l a -> l a
>>    decon :: l a
>>	     -> w -- on empty
>>	     -> (a -> l a -> w)
>>	     -> w
>>
>>instance List [] where
>>    nil = []
>>    cons = (:)
>>    decon [] on_nil on_cons = on_nil
>>    decon (h:t) on_nil on_cons = on_cons h t
>>

Yes. That 'decon' function corresponds to the 'case' function I went on 
to suggest later in my post, as part of a 
data-constructors-as-class-members language extension.

I used 'class List l a | l -> a', not 'class List l', so that there 
could be an instance for PackedString.

>we can then write truly generic list-processing functions:
>
>>rev:: List l => l a -> l a
>>rev l = rev' l nil
>> where
>>   rev' l acc = decon l acc (\h t -> rev' t (cons h acc))
>>
>
>not in a terribly convenient way though.
>
>   *Foo> rev "abc"
>   "cba"
>

It could get more convenient, though, if we have 
data-constructors-as-class-members and move the familiar list 
constructors into the List class. (The following might need to be built 
into the compiler, because of the special [] syntax.)

    class List l a | l -> a where
        [] :: l
        (:) :: a -> l -> l

    data OrthodoxLinkedList a = NilOLL | ConsOLL a (OrthodoxLinkedList a)

    instance List (OrthodoxLinkedList a) a where ...

Then existing definitions of list functions would work at the more 
generic level.

Regards,
Tom