mfix as an effect

[Ashley Yakeley]

In article <20040112005508.0B9DEAB8D at Adric.metnet.navy.mil>,
 oleg at pobox.com wrote:

> That is, Ashley Yakeley claimed that his Haskell implementation of
> Scheme implements letrec via Y -- and yet his implementation is
> consistent with the R5RS semantics, which mandates the updateable-cell
> letrec.

I am in any case withdrawing this claim, as it turns out my 
implementation of mfix uses IORefs explicitly.

-- 
Ashley Yakeley, Seattle WA