Coercing Variables

Question

Hello;
I have written a class that replaces strings in strings iteratively throughout the entire main string. However, when I call it, of course, it becomes and [object ReplaceString] when I really need it to be a plain old String. How do I get from one to the other?
Thanks,
Terry

Answer 1

Hello
Well, every now and then I see this sort of question.
return 1.1 - 1.0 = 0.1 -- > false
return 0.5 - 0.4 = 0.1 -- > false
return 0.7 - 0.3 = 0.4 -- > false
return 1.3 - 1.2 = 0.1 -- > false
etc.
Why?
Because it is the nature of binary floating point arithmetic of finite precision, that most computer uses in calculating decimal fractions and numbers larger than N-bit integer. (N depends upon architecture and library)
Indeed,
1.1 - 1.0 - 0.1 = 8.32667268468867E-17
0.5 - 0.4 - 0.1 = -2.77555756156289E-17
0.7 - 0.3 - 0.4 = -5.55111512312578E-17
1.3 - 1.2 - 0.1 = 8.32667268468867E-17
etc.
These results are 'correct', that is they conform to the rule of binary floating point arithmetic of double-precision (IEEE Standard 754).
cf.
Inside Machintosh: PowerPC Numerics (archived pdfs - one file per chapter)
http://developer.apple.com/documentation/mac/PPCNumerics/PPCNumerics-2.html
Idem. (one pdf file)
http://developer.apple.com/documentation/Performance/Conceptual/MacOSX_Numerics/Mac_OSXNumerics.pdf
(Yes, old document but still quite valid about IEEE 754 etc.)
So it is very important practice to round floating point number to some significant digits before comparison. It's also true in most programming language. Remember to round decimals before comparison!
Something like this.
--CODE
set x to 1.1 - 1.0
set y to 0.1
--return roundn(x, -1) = roundn(y, -1) --> true
--return roundn(x, -12) = roundn(y, -12) --> true
return roundn(x - y, -12) = 0 --> true
on roundn(x, d)
number x : source number
integer d : rounding target place (i.e. rounding up/down to (10 ^ d)'s place)
set h to 10 ^ d
return (round (x / h) rounding to nearest) * h
end roundn
--END OF CODE
Hope this may help,
H