Topes

["J. A. Durieux" <xerox@cs.vu.nl>]

(1) Yes, I realise that topes are just reified locations, i.e. something
    old and venerable.  I had understood that backward compatibility was a
    major reason not to accept changes in Scheme.  Therefore I introduced
    them in a way that did not break existing programs.
    From my biased point of view, I can only see advantages in having
    topes, so I wonder why they aren't there.
    BTW, "tope" is simply the Englified "topos", which is Greek for the
    Latin "locus", whence the English "location".  I felt that CS could use
    some fresh terminology instead of "re-using" "overloaded" terms. :-)

(2) I also understand that "load" is not required to be tail-recursive, but
    since its return value is unspecified, it is not precluded from being
    that either.
    But *is* there a way to catch a continuation that does not point back
    to the file?
    On the notion of tail-recursion: PC-Scheme 4.0 runs out of memory on
    ((call/cc call/cc) (call/cc call/cc)).  Is this a violation of the
    requirements? It doesn't have to eat memory (my interpreter loops in
    constant memory on this).

    Well, thanks for the responses so far..