Thomas Bushnell BSG wrote:
> Where is the Scheme function which modifies closures?

set!

as in:

(let ((i 0))
   (lambda ()
     (set! i (+ i 1))
     i))

> I was always taught that an atom is anything which is not a cons.  :)
> At least, the Little Lisper says so.

You may need to update your reading list ...

> But that aside, NaNs could also have structure, as has been pointed
> out;

Not if we're talking abou NaNs in the IEEE sense.

 > it could be useful for them to come with a tag identifying their
 > origin.

Perhaps, but let's concentrate on getting IEEE numerics right, and not
invent new ways of doing numerics.  At least not for R6RS.
--
	--Per Bothner
xxxxxx@bothner.com   http://per.bothner.com/