> > The Zipper technique is the generalization
> > of this idea to arbitrary algebraic datatypes; it can be generalized
to
> > higher-order data too (continuations).
I was too optimistic when I wrote this; someone actually e-mailed me asking
for a fuller explanation after I posted this, but I had to take this claim
back. (Unfortunately, I didn't correct myself in public.) At the time I had
been fooling around with the idea, but never figured out exactly how to
implement it.
> Is there some code around that illustrates the use of Zippers? They
> sound very interesting.
I have never seen anyone use the technique in any distributed Haskell code.
Gerard Huet wrote the article that appeared in JFP, and I believe he
remarked that they had used the technique to represent ASTs in a structural
editor (perhaps in the Coq distribution?).
--FAC