folding & knil placement Taylor Campbell 20 Dec 2003 17:44 UTC

I've seen various complaints about the placement of the KNIL in the
folding operations: in SRFIs 1 & 13, and currently 43 as well, the KNIL
argument is _after_ all the element arguments -- _even_if_ there may be
an arbitrary number of them.  Irritating, unnecessary, and easily
inefficient argument list destructuring with LAST, some variant of
SPLIT-AT, or something like that is the result of placing the KNIL at
the last argument place.  Though it would break compatibility, it would
be a great improvement otherwise; would anyone be opposed to moving
the placement of KNIL in the argument list of KONS using VECTOR-FOLD
and VECTOR-FOLD-RIGHT to the first argument, and if so, what's your
rationale for leaving it where it is?