nick rothwell | project cassiel

Here's a little problemette I'm having: when I pass a float like 

"4.53" into JavaScript from the Max world, it gets turned into 4.53 

+noise digits. Internally as a constant in JavaScript, it's fine. 

I assume the noise digits are due to rounding (the number 4.5 is 

fine, presumably because it has a precise IEEE representation). I 

don't know why it works internally in JS: perhaps that's doing some 

So, my question is: how do I truncate or round the accuracy of 

rendering floats into strings in JS? I couldn't find anything in the 

documentation regarding floating point formatting a la sprintf().

    post("arg gives: " + arguments[0] + "n");

    post("literal gives: " + 4.53 + "n");

Here's a little problemette I'm having: when I pass a float like  
"4.53" into JavaScript from the Max world, it gets turned into 4.53 
+noise digits. Internally as a constant in JavaScript, it's fine.  
(Example enclosed.)

I assume the noise digits are due to rounding (the number 4.5 is  
fine, presumably because it has a precise IEEE representation). I  
don't know why it works internally in JS: perhaps that's doing some  
kind of symbolic interpretation.

So, my question is: how do I truncate or round the accuracy of  
rendering floats into strings in JS? I couldn't find anything in the  
documentation regarding floating point formatting a la sprintf().

function foo() {
	post("arg gives: " + arguments[0] + "n");
	post("literal gives: " + 4.53 + "n");
}

Nick Rothwell / Cassiel.com Limited
www.cassiel.com
www.myspace.com/cassieldotcom
www.loadbang.net


compensating-for-loss-of-float-accuracy-in-javascript

On 30 juin 07, at 19:45, Nick Rothwell wrote:

> So, my question is: how do I truncate or round the accuracy of 

> rendering floats into strings in JS? I couldn't find anything in 

> the documentation regarding floating point formatting a la sprintf().

Somthing like that should work (no tested though):

> So, my question is: how do I truncate or round the accuracy of  
> rendering floats into strings in JS? I couldn't find anything in  
> the documentation regarding floating point formatting a la sprintf().

function bang()
{
	var toto = 0.576528356;
	post(toto.toFixed(3));
}

On 30 juin 07, at 20:33, Pierre Alexandre Tremblay wrote:

> I always wonder: is it really a lost of resolution, or is it the 

> opposite: we just see a 32 bit float being properly represented in 

JS works with 64 bit double float. but for representing the number as 

a string you usually need to display it with less precision.

> I always wonder: is it really a lost of resolution, or is it the  
> opposite: we just see a 32 bit float being properly represented in  
> 64-bit double float?

JS works with 64 bit double float. but for representing the number as  
a string you usually need to display it with less precision.

On 30 Jun 2007, at 18:55, Emmanuel Jourdan wrote:

Cool - thanks. I'm sure I didn't come across this when I dug through 

> 	var toto = 0.576528356;
> 	post(toto.toFixed(3));

Cool - thanks. I'm sure I didn't come across this when I dug through  
the JS 1.5 docs I have here...

Moving this thread to the javascript forum, please direct your replies to the javascript list.

Compensating for loss of float accuracy in JavaScript?