Thorsten Froehlich <tho### [at] trfde> wrote:
: Do you really think there is some code in there that says:
: "add 0.00000000298023224 to clock to annoy users"?  ;-)

  I was wondering if floats could *really* be that inaccurate. When I make
this with povray:

#debug concat(str(99/100,0,25), "\n")

what I get is:

0.9899999999999999911182158

(instead of 0.99). I though, "is this really possible?".
  Then I wanted to be sure and I made a small C program which does basically
the same thing:

#include "stdio.h"
int main()
{
    printf("%.25f\n", 99.0/100.0);
    return 0;
}

  This program prints, surprise surprise:

0.9899999999999999911182158

  So now I am convinced that there's no bug in povray.

  Of course, when making a scene file, one has to minimize the effects of
floating point inaccuracy.
  For example, don't do it this way:

#declare Ind = 0;
#declare Clock = 0;
#while(Ind <= 100)
  #debug concat("Clock: ", str(Clock,0,25), "\n")
  #declare Clock = Clock+1/100;
  #declare Ind = Ind+1;
#end

(The last number printed is 1.0000000000000006661338148.)
  Instead, do it this way:

#declare Ind = 0;
#while(Ind <= 100)
  #declare Clock = Ind/100;
  #debug concat("Clock: ", str(Clock,0,25), "\n")
  #declare Ind = Ind+1;
#end

(The last number printed is 1.0000000000000000000000000.)

  The reason for the second code being better than the first is left as
an exercise to the reader.

-- 
#macro M(A,N,D,L)plane{-z,-9pigment{mandel L*9translate N color_map{[0rgb x]
[1rgb 9]}scale<D,D*3D>*1e3}rotate y*A*8}#end M(-3<1.206434.28623>70,7)M(
-1<.7438.1795>1,20)M(1<.77595.13699>30,20)M(3<.75923.07145>80,99)// - Warp -

Post a reply to this message