>
> > This sort of thing is also handy for collections like circular lists,
> > wrap-around vectors, and similar collections.
> Maybe, maybe not.  The semantics of collection-fold are such that
> an enumeration over a collection should enumerate once over all the
> values in the collection unless halted by the folding function.
>
> Circular lists and simliar collections can be created using the
> negative index trick hinted at by the SRFI, but I'm unsure whether
> folds should proceed indefinitely over finite collections.  Other
> thoughts?

Actually, I'm vehemently against collection-fold ever looping over a
circular but finite collection.  Half of the advantage of collection
fold is that it allows you to pass functions unaware of the properties
of the underlying collection and let the implementation handle
retrieving all the requisite values.

Potentially or actually infinite collections are another matter.  Its
reasonable to pass a function to fold which won't terminate an infinite
stream, for example, passing a function to display characters from a
TCP/IP stream.

	Scott