Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!uakari.primate.wisc.edu!aplcen!haven!adm!cmcl2!lanl!jlg
From: jlg@lanl.gov (Jim Giles)
Newsgroups: comp.lang.fortran
Subject: Re: rounding a real to a whole number
Message-ID: <5524@lanl.gov>
Date: 10 Nov 90 02:04:06 GMT
References: <6125@harrier.ukc.ac.uk>
Organization: Los Alamos Natl Lab, Los Alamos, N.M.
Lines: 52

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.

How about:

      Subroutine ticks(lbound,hbound,bottom,top,tick)
C       takes low and high bound of interval and returns
C       a 'tick' value for the interval between tick marks
C       on a graph axis.  The value of tick is always
C       either a power of 10, twice a power of ten or
C       five times a power of ten.  A new interval is
C       also returned which begins below the original
C       and ends above the original and is an integral
C       number of 'tick's long.  The number of ticks in
C       the new interval is (about) 5 to 13.
      real lbound,hbound,bottom,top,tick
      real lowtic, basetk
      integer tenp
      lowtic = (hbound-lbound)/12.5
      tenp=int(alog10(lowtick))
      if (tenp.gt.alog10(lowtick)) tenp=tenp-1
      basetk = 10**tenp
      if (tick.lt.lowtick) tick = 2 * basetk
      if (tick.lt.lowtick) tick = 5 * basetk
      if (tick.lt.lowtick) tick = 10 * basetk
      bottom = nint(lbound/tick)*tick
      if (bottom.gt.lbound) bottom = bottom - tick
      top = nint(hbound/tick)*tick
      if (top.lt.hbound) top = top + tick
      return

Given your numbers: -123.445 -> 1033.02, this routine returns
TICK equal to 100.0, BOTTOM equal to -200.0, and TOP equal to 1100.0.

The number of ticks in the result interval is always less than the
divisor for LOWTICK plus one (in this case, the divisor is 12.5,
so the maximum number of ticks is 13).  The minimum number of ticks
in the final interval is the LOWTICK divisor itself divided by 2.5.
So, if you want finer resolution, pick a number larger than 12.5.
If you want coarse resolution, pick a smaller number to divide by.
You can even make the divisor one of the subroutine arguments if
you want.  Given your numbers:-123.445 -> 1033.02, and a divisor
of 25, the answer is - TICK = 50.0, BOTTOM = -150.0, and TOP = 1050.0
giving 24 intervals.

Not elegant, but not bad off the top of my head.

J. Giles