NewsForumsCodeLogsGallerySpecsSearch

Matematical question

Monkey Forums/Monkey Programming/Matematical question

Vinians(Posted 2011) [#1]
Hi, I need to solve a simple question.
I'm creating a funcion to snap a instance like it:
snape_instance(32, 32)

With this code, if X was 50 it should go to 64 not to 32 because the instance is near to 64 than 32. How can I create a funcion like that? Or
how can I get the decimal part of a float? Like 50 / 32 = 1.56 I want the .56 part. But please only using matematics not strings.
My solution was this:
	Method MoveSnap(sx%, sy%)
		Local dx#, dy#
		dx = (x / sx)
		dx = Abs(Floor(dx) - dx)
		If (dx > 0.5) 
			x = (Floor(x / sx) + 1) * sx
		Else
			x = (Floor(x / sx)) * sx
		End
		dy = (y / sy)
		dy = Abs(Floor(dy) - dy)
		If (dy > 0.5) 
			y = (Floor(y / sy) + 1) * sy
		Else
			y = (Floor(y / sy)) * sy
		End
	End

But its so slow, maybe another solutions?


GfK(Posted 2011) [#2]
Try x = ceil(x / 32.0) * 32


Vinians(Posted 2011) [#3]
in your example fails because he always rounds up. That's the problem. I need it to always round to the nearest multiple of 32, which can be above or below.


GfK(Posted 2011) [#4]
I'm on my android phone atm so i can't test


Warpy(Posted 2011) [#5]
X=ceil(x/32. 0-. 5) *32

(also written on my android phone, so not tested)


Vinians(Posted 2011) [#6]
buddy I do not understand your code


GfK(Posted 2011) [#7]
He formatted it a bit funny but it seems to work:
x= Ceil(x/32.0 - 0.5) *32



Vinians(Posted 2011) [#8]
@Warpy Humm.. Thats nice! This reallt works, but how do you discover it?
@Gfk thanks to translate :D


Roger Lockerbie(Posted 2011) [#9]
Subtracting point five solves your rounding up issue. It's a common idiom


FlameDuck(Posted 2011) [#10]
Why no love for Floor?!?


Warpy(Posted 2011) [#11]
man, I just used Ceil because that's what gfk used.

Vinians: here's why my code works:

Ceil rounds any number up to the nearest whole number. So anything in the range N up to N+1 is rounded to N+1.

Rounding to the nearest number works like so: if the number is less than N+0.5, the number is rounded down to N, and if the decimal part is greater than or equal to N+0.5, the number is rounded up to N+1.

So we want everything in the range N-0.5 up to N+0.5 to be rounded to N. If you subtract 0.5 from those numbers, you get the range N-1 up to N. Calling Ceil on those numbers will round them up to N.

So Ceil(x-0.5) rounds x to the nearest whole number.

But you want to round to the nearest multiple of 32. So for a number x, think about what multiple of 32 it is and round that.

What I mean is, think about x = y*32. Divide by 32 to get y = x/32.0. Now round y to the nearest whole number ( y = Ceil(y-0.5) ), and multiply by 32 to get the nearest multiple of 32 to x.

So the closest multiple of 32 to x is Ceil(x/32.0 - 0.5)*32.

I hope this explanation helps.