 |
Author |
Topic: permanent decimal spot rounding? (Read 730 times) |
|
lunetta
 
 
|
permanent decimal spot rounding?
« on: Mar 1st, 2005, 7:02pm » |
|
Hello all
I'm wondering if there's a painless way to determine that all numbers everywhere in my sketch will have, lets say 5 decimal spots. So, all numbers above that should be rounded with a floor value...
example: 5.1234567890234 would become 5.12346...
What I have done is a function
float rounder (float number)
{
return floor(number*100000)/100000;
}
but the problem is:
- it only works with floats
- it has to be called everywhere...
This is not a problem per se, but a curiosity if I can define a number datatype just like if I were extending a built in class...
|
|
|
|
|
fry
|
Re: permanent decimal spot rounding?
« Reply #1 on: Mar 1st, 2005, 9:07pm » |
|
you'd need to create some sort of a clever class and use that. the general method would be that you'd use an int (or a long) to store the numbers, and then pull em out as needed. there isn't really a way to extend the primitives (int, float, etc) that are allowed in processing.
then again, if all you need is formatting for the user, nf() is a simple way to get decimals formatted cleanly into a string:
http://processing.org/reference/nf_.html
|
|
|
|
|
lunetta
|
Re: permanent decimal spot rounding?
« Reply #2 on: Mar 1st, 2005, 10:14pm » |
|
great, thanks for your response.
What I'm in need is to control the aberrations due to small numbers float imprecision, but I guess I'll keep using ints as much as possible and convert them to float only when necessary...
|
|
|
|
|
fry
|
Re: permanent decimal spot rounding?
« Reply #3 on: Mar 2nd, 2005, 3:52am » |
|
you might also consider "fixed point" numbers, which are integer based, and might give you more straightforward results... a quick google or wikipedia search should explain how they work.
|
|
|
|
|
st33d
|
Re: permanent decimal spot rounding?
« Reply #4 on: Mar 2nd, 2005, 9:00pm » |
|
I wondered about this myself and typed a load of words into google: float decimal point length java
http://www.prosperosoftware.co.uk/software/java/decimals/uk/co/prosperos oftware/pspnum/PSPvFloat.html
So if I've got this right, the floats only go up to six digits anyway. One less than six. To store a really fiddly number you would need a double, and then you need to trim it back to a float or an int to use it with the Processing functions (which don't seem to use doubles - thank God).
|
I could murder a pint.
|
|
|
lunetta
|
Re: permanent decimal spot rounding?
« Reply #5 on: Mar 3rd, 2005, 6:10am » |
|
wow, nice info, thanks
|
|
|
|
|
toxi
|
Re: permanent decimal spot rounding?
« Reply #6 on: Mar 3rd, 2005, 11:27am » |
|
also posted a message (incl. examples) about fixed point numbers a (long) while ago over in this thread. it shouldn't be hard to wrap this in a class for use as custom datatype...
|
http://toxi.co.uk/
|
|
|
|
