Formal comment #26 (enhancement)

Map and for-each should work even if lists are of unequal length
Reported by:	John Cowan

Component:	baselibrary
Version:	5.91

The R6RS definitions of "map" and "for-each" should permit the list
arguments to be of unequal length, and specify that they terminate
when the shortest list runs out. This is what SRFI-1 does, it is
cheap, and it is useful.

RESPONSE:

The proposal is indeed cheap to implement, may be more useful
than the draft R6RS semantics, and has one of the most
widely used SRFIs as precedent.  It may, however, mask errors
where lists of different lengths are supplied by mistake, and
the semantics becomes unclear when input lists are extended
destrutively as side effects of invoking the procedural arguments.

To investigate precedent further, several implementations
were tested to see whether they allow list arguments of
different lengths:

Bigloo 2.7a             error
Petite Chez 7.0a        error
Chicken 2,5             error
Gambit-C v4.0b20        error
Kawa 1.8                okay (but does not allow circular lists)
Larceny v0.93           okay (but does not allow circular lists)
MIT Scheme 7.7.90.+     error
MzScheme v352           error
Scheme 48 v1.3          okay (but does not allow circular lists)
SCM 5e1                 okay (even with circular lists)

It is questionable whether the SRFI-1 extension of these
procedures to accept lists of differing lengths is so
widely accepted that it should be required by the R6RS.
On the other hand, the SRFI-1 extension is common enough
the R6RS should perhaps not preclude it.

One solution would be to specify map and for-each essentially
as in R5RS:  If list arguments have different lengths, then
an exception might be raised, but is not required.
Portable programs would not be able to depend upon the
extension, and would also not be able to depend upon the
exception being raised.

While we are still studying this issue, the editors intend to
resolve this issue in a manner consistent with our resolution
of formal comment #87.