On Thu, Jan 8, 2015 at 4:02 AM, Keean Schupke <[email protected]> wrote:

> Thinking about:
>
> instance Ord int is
>     x < y = do
>         count <- count + 1
>         primative_int_less(x, y)
>
> Its not nice that the instance is gobal, and that implies count is global.
>
Globals aren't always evil. Yet this instance is perfectly valid, and it
can "instantiated through" in the way I have been talking about without any
difficulty.

And while I agree with you that perhaps this case should be a procedure,
that isn't always possible. In this case we might have had a pre-existing
arrangement of code that we are temporarily instrumenting.

But the main point I was making is that we obviously can't think of a thing
with state as something we can lift into a type.


shap
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to