Re^2: Is this odd behavior a floating point problem?

Replies are listed 'Best First'.
Re^3: Is this odd behavior a floating point problem? by davido (Cardinal) on Mar 23, 2012 at 18:09 UTC
It's not a Perl problem, it's the problem of representing a non-terminating series for base 2 using a finite number of binary digits. You are accustomed to seeing it in base ten when you try to represent 1/3rd, yet I hear no complaints that your ten counting fingers are malfunctioning. Your calculator gets it right by rounding to the eight or ten significant digits that you see on its little LCD display. In other words, it really doesn't get it right; it just covers up the ugliness. And, in fact, I see a nearly identical question every day in reference to C, C++, PHP, and myriad other programming languages over on StackOverflow. It's not a problem unique to Perl. I understand that the "What every computer scientist should know..." article is a little beyond what someone who just wants to get the job done might want to digest. That's fine, the article goes into painful details. Try this response (shameless plug, I wrote it), which tries to spell it out in less technical terms: Re: shocking imprecision. Dave	[reply]
Re^3: Is this odd behavior a floating point problem? by BrowserUk (Patriarch) on Mar 23, 2012 at 19:13 UTC
If you don't like the WhatEvery... article -- and I agree that it is written more to impress than inform; and is way over promoted -- then try this. Short, sweet & clear. With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday' Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". In the absence of evidence, opinion is indistinguishable from prejudice. The start of some sanity?	[reply]
Re^3: Is this odd behavior a floating point problem? by Anonymous Monk on Mar 23, 2012 at 19:38 UTC
Yes, I've seen and skimmed this document, but, not to be rude, but I'm not interested in becoming a computer scientist in order to write a script to do basic math. Adding together 0.001 40 times is pretty basic and if my calculator can do it, I not sure I understand why Perl won't. :) Try site:perlmonks.org What Every Computer Scientist Should Know About Floating-Point Arithmetic and you can learn from others who weren't satisfied with that document Simple math gone wrong Floating Point Looping In Perl Math 101 anyone? adding numbers and floating point errors sprintf rounding convention perl subtraction RFC: Large Floating Point Numbers - Rounding Errors perl subtraction Using (s)printf() http://en.wikipedia.org/wiki/Floating_point#Addition_and_subtraction limit of for loop!! Why does perl math suck? Perl 5.8.x floating point representation error Strange rounding Error in Perl Bug? 1+1 != 2 [Win32, C, and way OT] C floats, doubles, and their equivalence My floating point comparison does not work. Why ? cos (100000000.0) Floating point problems Ever heard of pi? 22/7? Guess what, 22/7 exceeds p General Decimal Arithmetic Decimal Arithmetic FAQ(Frequently Asked Questions) Now you say you're studying chemical diffusion so I assume you would have heard of significant figures? Surely your professor, when discussing significant figures, would have explained the basic limitations of adding machines (calculators/computers)? I was hoping, after reading that document, you would ask explicitly how to round numbers for display purposes in perl. perlfaq4#Does Perl have a round() function? What about ceil() and floor()? Trig functions? Math::BigRat Math::BigInt Math::BigFloat bigint/bigrat/bignum While you can create a calculator using perl (like Tk::Calculator::RPN::HP ), and expect it to do rounding like your pocket calculator, perl itself, not being a calculator, won't hide the details of floating point arithmetic from you, so it is good knowledge to have. Any scientist using computers for calculations needs to know the limits of his tools.	[reply]
Re^4: Is this odd behavior a floating point problem? by Anonymous Monk on Aug 22, 2012 at 13:10 UTC
See also What Every Programmer Should Know About Floating-Point Arithmetic or Why don’t my numbers add up? discovered via MidLifeXis	[reply]
Re^3: Is this odd behavior a floating point problem? by ww (Archbishop) on Mar 24, 2012 at 00:29 UTC
If I were doing it with a calculator, I'd surely end up with too many or too few entries of 0.001. I might mess it up building an array. Are you sure you have the right number of elements? You can check easily enough: `say scalar(@arr); # 40 say $#arr; # 39` [download]	[reply] [d/l]


Perl Monk, Perl Meditation
	PerlMonks