Bioequivalence and Bioavailability Forum

Binary representation [Software]

posted by ElMaestro  – Denmark, 2018-03-16 14:13 (1837 d 03:53 ago) – Posting: # 18561
Views: 6,015

Hi Mittyri,

❝ x=0.250000

❝ as it should be

❝

How nice, then it is also compiler-dependent...
The issue is that 0.05 cannot be represented accurately in a binary brain.
So, the compiler will need to deal with such a fraction and either round it up or round it down. Whatever it does, it can have downstream implications. In my case it implied that my simple naïve decimal logic (0.25 <=0.25) never evaluated to TRUE because it was not tested as such.
The solution could be to add e.g. 1e-8 to the righthand side. Dumb, but that's logic.
Computers suck.

Complete thread:

• Nervous ticks in C - a reminder of the cruel logic of logic ElMaestro 2018-03-15 10:24 [Software] 
  • Nervous ticks in C - a reminder of the cruel logic of logic mittyri 2018-03-15 11:55
    • Binary representationElMaestro 2018-03-16 13:13
      • bypassing eqaulity conditions mittyri 2018-03-17 18:12
        • And R? Helmut 2018-03-18 00:08
          • And R? ElMaestro 2018-03-18 08:29
  • Nervous ticks in C - a reminder of the cruel logic of logic nobody 2018-03-18 13:13
    • Nervous ticks in C - a reminder of the cruel logic of logic ElMaestro 2018-03-19 09:44
      • Nervous ticks in C - a reminder of the cruel logic of logic nobody 2018-03-22 08:56
        • Check it out for yourself - Dissolution Bootstrap:-) ElMaestro 2018-03-22 17:16
          • Check it out for yourself - Dissolution Bootstrap:-) Helmut 2018-03-22 18:50
            • Check it out for yourself - Dissolution Bootstrap:-) ElMaestro 2018-03-22 19:04
              • Check it out for yourself - Dissolution Bootstrap:-) Helmut 2018-03-22 22:22