formula for modulo?

[Argitoth]

need floating point modulo, so need formula for jsfx rather than using built in one.

anyone know it? thanks!

got it

function mod(f,mod) ( f-floor(f/mod)*mod );

[SaulT]

How is that different from using the % operator?

[Tale]

The % operator works with integer numbers, while Argitoth's mod() works with floating point numbers. And I think they return different signs when the modulus is negative.

[planetnine]

I define my own INT and MOD functions in EEL -INT is floor for >0 and ceil for <=0, and then MOD uses INT.



>

[Banned]

Originally Posted by planetnine

I define my own INT and MOD functions in EEL -INT is floor for >0 and ceil for <=0, and then MOD uses INT.

That makes sense for 'better' rounding of a float value to an int, but for a floating point modulo function, one may also need the fractional part of the remainder after division by an integer, no?

[planetnine]

Well it's easy enough to make a REM function in addition to your MOD function, once you have a functioning INT

function int(val) local(res)
(
	val >=0 ? // calculates true integer of pos and neg values
	(
		res = floor(val);
	):(
		res = ceil(val);
	);
	res;
);


function mod(num, divisor)
(
	int(num/divisor);
);

function remdr(num, divisor)
(
	num - (mod(num) * divisor);
);

Edit: Hehe, I'm using modulo for the integer part and remainder for what is actually modulo. If you ignore the screwed-up terminology, it works



>

[Banned]

Is it? I don't see why conversion to int is a prerequisite for a MOD function. After conversion to int, you have lost the fractional part. So I'd rather just have a MOD function that works with float input and output, and convert to int when needed afterwards.

[planetnine]

If you read my Edit, I've used the term mod erroneously for the integer division part, and "remainder" for the actual mod value.

I can't remember the mathematical term for the integer division bit

So, more like:

function int(val) local(res)
(
	val >=0 ? // calculates true integer of pos and neg values
	(
		res = floor(val);
	):(
		res = ceil(val);
	);
	res;
);


function int_div(num, divisor)
(
	int(num/divisor);
);

function mod(num, divisor)
(
	num - (intdiv(num, divisor) * divisor);
);

It can be written more compactly, but it depends if you need the intermediate stages' functions in your script.



>

[SaulT]

Are we saying that JS converts floats to integers before doing a modulus?

[Tale]

Originally Posted by SaulT

Are we saying that JS converts floats to integers before doing a modulus?

Well, if you use the % operator then: Yeah, I guess. C/C++ also does this, the reason is probably that the % operator results in a DIV instruction.