Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!van-bc!ubc-cs!cheddar.ucs.ubc.ca!buckland
From: buckland@cheddar.ucs.ubc.ca (Tony Buckland)
Newsgroups: comp.lang.fortran
Subject: Re: rounding a real to a whole number
Message-ID: <10292@ubc-cs.UUCP>
Date: 1 Nov 90 19:28:00 GMT
References: <U714092.90Nov1101629@eagle.larc.nasa.gov>
Sender: news@cs.ubc.ca
Reply-To: buckland@cheddar.ucs.ubc.ca (Tony Buckland)
Organization: UBC Computing Centre, Vancouver, B.C., Canada
Lines: 30

In article <U714092.90Nov1101629@eagle.larc.nasa.gov> u714092@eagle.larc.nasa.gov (prichard devon ) writes:
|
|what I want to do is to have a subroutine which, given a range of real
|numbers, for example  -123.445 -> 1033.02 , and choose whole numbers 
|outside that range. obviously the first application is axis values for
|x-y plots.  optionally, picking a 10-multiple of whole numbers ought to
|be included; for  -1.45567 -> 3.34112 , it should return -1.4 and 3.4 ...
|
|while this seems so simple, I don't see a way to do this reasonably with
|Fortran intrinsics.  I just can't get it to work correctly with truncations.
|
|anyone done this sort of thing listening out there?

 I imagine you meant -1.5, not -1.4.
 That quibble aside, a traditional way of rounding up is to add
 a floating-point quantity before truncating to integer.  Watch
 what happens in this example:
 
    f = 3.34112                    unrounded quantity
    f = f + 0.09999                add some variation on "just
                                   less than one"
    f = 10.0 * f                   we want to keep two digits
                                   f is now 34.4111
    i = f                          i is now 34
    f = i                          f is now 34.0000
    f = f / 10.0                   f is now 3.4
 
 The negative case is left as an exercise for the student.
 The problem with inexact representation of floating-point
 machine representation of most non-integers is ignored.