JavaScript Woes

The problem is "a lack of precision"

While JavaScript is quickly becoming the lingua franca of the web, this general purpose language contains built-in limitations which restrict its use to simple applications. For example, I recently tried to write a Diffie-Hellman Demo in JavaScript but detected problems during testing which I traced to limited math precision. Check out these comparisons:

• Many junior programmers think this problem is related to internal rounding but it is actually a limitation due to float-point math.
• Many seasoned programmers are surprised to learn that all JavaScript numbers (floating point or not) are stored in 64-bit floating point format (similar to JAVA's double) and there are no official alternative data types.
• References:
  • http://www.hunlock.com/blogs/The_Complete_Javascript_Number_Reference
  • http://www.crockford.com/javascript/survey.html
• JavaScript workarounds:
  • https://github.com/jhs/bigdecimal.js/ (Big Decimal)
  • https://github.com/silentmatt/javascript-biginteger/ (Big Integer)
  • http://www.leemon.com/crypto/BigInt.html (Big Int)
  • http://www-cs-students.stanford.edu/~tjw/jsbn/ (Big Number)
• Other workarounds:
  • for client side software, switch to JAVA
  • for server side software, switch to PERL

PERL Workaround

Legend:	<sr> = system response
	<ur> = user response

<sr>	$
<ur>	type bigint_demo.pl
	#
	#       title: bigint_demo.pl
	#
	use bigint;             # use bigint after this point
	$x = 7 ** 19;           # x = 7^19
	printf("%s\n",$x);      # "%s" because other formats bomb
<sr>	$
<ur>	perl bigint_demo.pl
<sr>	11398895185373143	! correct
<ur>	$

Related Links

• Neil's OpenVMS Notes

Back to OpenVMS
Back to Home
Neil Rieck
Kitchener - Waterloo - Cambridge, Ontario, Canada.