Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site digi-g.UUCP
Path: utzoo!linus!decvax!tektronix!uw-beaver!cornell!vax135!houxm!ihnp4!stolaf!umn-cs!digi-g!brian
From: brian@digi-g.UUCP (Brian Westley)
Newsgroups: net.lang
Subject: Summary of modulus function
Message-ID: <220@digi-g.UUCP>
Date: Thu, 6-Sep-84 12:14:57 EDT
Article-I.D.: digi-g.220
Posted: Thu Sep  6 12:14:57 1984
Date-Received: Sun, 16-Sep-84 12:24:58 EDT
Organization: DigiGraphic Systems Corp., Mpls.  MN
Lines: 22

Here is a summary of the responses I received about the 'C' & Pascal &
Modula, etc. modulus functions.

I asked if (A mod B) should always fall within 0..B-1 even with negative A.
It seems that the usual mathematical definition of modulus says yes, while
most programming languages say no.

Most programming languages really use remainder, such that:
(A div B) * B + (A mod B) = A    or:
(A mod B) = A - (A div B) * B
(This still depends on whether div truncates towards zero, or always down)
Anyway, most languages return -2 for (-2 mod 3), while math modulus would
say (-2 mod 3) is 1.

Also, in math modulus, if B is negative, the answer is in B+1..0
Some langs don't allow B (or even A) to be negative.

Wouldn't it be nice to have a modulus function, a remainder function,
and two integer division functions, one truncating towards zero, the other
always truncating to largest integer <= A/B ?
I'm tired of having to do ((A+B) mod B) to get modulus to work the way I
want it to....