Issue BUTLAST-NEGATIVE Writeup

Issue:        BUTLAST-NEGATIVE

Forum:	      Editorial

References:   BUTLAST (p271), NBUTLAST (p271)

Category:     CLARIFICATION/CHANGE

Edit history: 01-Mar-91, Version 1 by Pitman

Status:	      For X3J13 consideration


Problem Description:


  What happens if the second argument to BUTLAST and NBUTLAST is negative?


Proposal (BUTLAST:SHOULD-SIGNAL):


  Specify that if the second argument, N, to BUTLAST or NBUTLAST is negative, 

  then an error `should be signalled'.


Test Cases:


  1. (BUTLAST '(A B C) -2)


  2. (LET ((X (LIST 'A 'B 'C)))

       (NBUTLAST X -2))


Rationale:


  Mostly this precludes implementations from adopting other, 

  variant interpretations.


Current Practice:


  1. Symbolics Genera returns (A B C NIL NIL).


  2. Symbolics Genera signals an error.


Cost to Implementors:


  Small.


Cost to Users:


  None.  Portable programs cannot currently rely on any particular 

  behavior.


Cost of Non-Adoption:


  Users would be confused about what to expect because implementations might

  adopt creative interpretations (such as Genera's) which are not supported

  in other implementations and so become a barrier to portability.


Benefits:


  Tighter spec.


Aesthetics:


  Genera's alternate interpretation of BUTLAST is arguably aesthetic, but

  does not extend well to NBUTLAST.  The interpretation proposed here makes

  BUTLAST and NBUTLAST more consistent.


Discussion:


  Pitman supports this proposal.