[Haskell-cafe] Need ideas how to model the lack of something

[Patrick Redmond]

Replies inline.

On Sunday, December 13, 2015, martin <martin.drautzburg at web.de> wrote:

> Am 12/13/2015 um 11:57 PM schrieb Patrick Redmond:
> > I think you're off to a good start with this insert signature:
> >
> > insert :: a -> C a -> Maybe (C a)
> > "Insert element `a` into structure `C a` and return a new structure if
> the insertion was successful."
>
> This way I'd have to be explicit about what C really is, don't I? Data.Set
> certainly has a very explicit data structure
> under the hood. I was hoping to express the idea of "something that can be
> inserted to and removed from" without
> specifying how the data is actually stored.


Use a typeclass.

But maybe that's a bad point to start from. At least this is where the
> trouble started when I tried to implement
> something on top of it. I just didn't have enough "flesh" to work with.


Yes, you will have to write a concrete implementation anyway, so start with
that. Make an explicit data structure, with concretely typed functions to
manipulate it.

When you have two of these explicit implementations, make a typeclass and
provide two instances - one which delegates to each of the implementations.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20151213/8cc53342/attachment.html>