$ans = 1, but ($ans = 1) = False

[Mat]

I decided to jump into the deep end on project euler, and try 251. For those interested, a brute force method should work, just very slowly.

A triplet of positive integers (a,b,c) is called a Cardano Triplet if it satisfies the condition:

2, 1, 5 was the example they gave, yet I wasn't able to reproduce it using my formula. After some checking, it turns out the formula was returning 1, but when asked whether it = 1, it returns false. script:

$a = 2
$b = 1
$c = 5

$ans = 0
$ans = ((root ($a + ($b * sqrt ($c)), 3)) + (root ($a - ($b * sqrt ($c)), 3)))
MsgBox (0, $ans = 1, $ans)

Func Root($fNum, $nExp = 3)
   Local $bNeg = False, $fRet = 0
   If $nExp < 0 Then Return SetError (1, 0, $fNum)

   If $fNum < 0 Then ; is negative
      If Mod($nExp, 2) Then ; nExp is odd, so negative IS possible.
         $bNeg = True
         $fNum *= -1
      Else
         Return SetError(1, 0, $fNum & "i") ; Imaginary number.
      EndIf
   EndIf

   $fRet = $fNum ^ (1 / $nExp)
   If $bNeg Then $fRet *= -1
   Return $fRet
EndFunc ; ==> Root

$ans is shown as 1

$ans = 1 is false!!

Any ideas? as it is starting to bug me.

MDiesel

[Jos]

Flootingpoint calculation. try

$ans = Int((root ($a + ($b * sqrt ($c)), 3)) + (root ($a - ($b * sqrt ($c)), 3)))

[Mat]

ah! I knew floating point numbers did provide some hassle for us, but I was not aware that it would affect this, thanks jos!

Should have just about done it then!!!

MDiesel

[Mat]

Sorry... but thats wrong too

Int rounds (or negates any decimals), so int(1.1) = 1

That means I end up way off...

MDiesel

[Mat]

I can't edit the last post

Solution I think:

$ans = int (((root ($a + ($b * sqrt ($c)), 3)) + (root ($a - ($b * sqrt ($c)), 3))) * 10000)
If $ans = 10000 Then MsgBox (0, "", "")

I can also edit the accuracy...

MDiesel

[Jos]

Have you tried using the round() function to limit the number of decimal places?

[Mat]

?? The problem is the rounding !!! I don't want to limit the number of decimal places!

The thing is... Int rounds. I need it to equal exactly 1, not 1.1

My method does NOT work incidentally, i am trying to find out why. All I need is a way to accuratetll tell if floating point number X == 1.

It seems to be pretty annoying...

There must be some way...

MDiesel

[Jos]

?? The problem is the rounding !!! I don't want to limit the number of decimal places!

The thing is... Int rounds. I need it to equal exactly 1, not 1.1

My method does NOT work incidentally, i am trying to find out why. All I need is a way to accuratetll tell if floating point number X == 1.

It seems to be pretty annoying...

There must be some way...

MDiesel

I understand but you only have 15 decimal places to play with. doesn't rounding on 14 decimals work for you? Edited July 18, 2009 by Jos

[trancexx]

@mdiesel; Jos is right of course. It's a matter of how your processor (or any) interprets floating-point numbers.

Floating-point number is 32 bits in length (as int is). So all of it needs to be within those freaking 32 bits.

This means that two different numbers (in nature - and not only two) are likely to be represented as the same floats by the machine.

This is the cause of wrong results.

This also means that you need to calculate with divergence.

All this to say that rounding is essential when evaluating mathematical operations with floating-point numbers done by machine. Avoiding that is wrong.

[Mat]

So your saying Round (14) would do it?

at a friends house atm, so ill check later. but right now i'm not exactly thinking strait,

mdiesel

[Xand3r]

here's an easy fix for that

MsgBox (0, String($ans) = 1 , String($ans))

[Mat]

here's an easy fix for that

MsgBox (0, String($ans) = 1 , String($ans))

hmm... Thats a good idea too... I'm trying everything out at the moment.

MDiesel