home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!usc!rpi!ghost.dsi.unimi.it!univ-lyon1.fr!frmop11!psuvm!auvm!VNET.IBM.COM!MFC
- Message-ID: <REXXLIST%92121511453666@UGA.CC.UGA.EDU>
- Newsgroups: comp.lang.rexx
- Date: Tue, 15 Dec 1992 16:44:38 GMT
- Sender: REXX Programming discussion list <REXXLIST@UGA.BITNET>
- From: Mike Cowlishaw <mfc@VNET.IBM.COM>
- Subject: Rounding twice in format()
- Comments: To: rexxlist@uga.uga.edu
- Lines: 11
-
- Yes, this is 'by design' -- it's assumed that built-in functions are
- not 'privileged' operators, and so should simply round their inputs
- (if "over-precise") before use. The X3J18 (ANSI standard) committee
- favours a 'SIGNAL ON DIGITS' option that would let one detect when
- rounding was occurring in this kind of context.
-
- Even/odd rounding was discounted for REXX because it implies processing
- an arbitrary number of digits, even if NUMERIC DIGITS is known (see
- footnote 40 on page 130 of "The REXX Language").
-
- Mike Cowlishaw
-