F64 lib Updated.
Posted: Wed Jul 28, 2004 12:54 am
PureBasic Forums - English
http://www.purebasic.com
https://www.purebasic.fr/english/
F64 lib Updated.
Page 1 of 2
Posted: Wed Jul 28, 2004 1:09 am
Thank you Jack! Your efforts are greatly appreciated.
Re: F64 lib Updated.
Posted: Wed Jul 28, 2004 9:52 am
Don't be that decent, it's also your lib now. I only brought it to live and you are keeping it alive nowjack wrote:for those that use the F64 lib by freedimension
Great work btw. Couldn't figure out what was wrong on that bug (don't like C that much
)Posted: Wed Jul 28, 2004 11:06 am
you are too kind, as my contribution is rather small. 
Posted: Wed Jul 28, 2004 11:32 am
but nonetheless important to the communityjack wrote:you are too kind, as my contribution is rather small.
Posted: Wed Jul 28, 2004 11:50 am
thank you, btw i emailed the modified source code to you, compiles with MSVC 6.
Posted: Wed Sep 07, 2005 9:19 pm
Howdy,
Shouldn't...
Return 621.68 and not 621.67? I'm using F64 version 0.95.
Great tool, by the way. Very very useful for me in many apps
Shouldn't...
Code: Select all
a.double
b.double
F64_Val(a, "621.675000")
F64_Round(a, 2, b)
Debug F64_Str(b)Great tool, by the way. Very very useful for me in many apps
Posted: Wed Sep 07, 2005 10:01 pm
Hi Jack and thanx for updating this excellent lib wich i can't live without (really).
Keep up the good work
Keep up the good work
Posted: Thu Sep 08, 2005 12:19 am
the function F64_Round rounds the value up, so the result is correct.Xombie wrote: Shouldn't...
Return 621.68 and not 621.67?Code: Select all
a.double b.double F64_Val(a, "621.675000") F64_Round(a, 2, b) Debug F64_Str(b)
thanks KarLKoX and Xombie for the compliments
Posted: Thu Sep 08, 2005 3:11 am
That's what I mean. 621.675 when rounding to two decimal places the 5 should round 7 up to 8 right? So it should be 621.68, right? The code I posted returns 621.67 as if it's just returning two decimal places and not rounding at all.
Posted: Thu Sep 08, 2005 9:29 am
Sorry I misunderstood Xombie.
not sure it’s a bug, it seems to be standard practice to round-off if the fractional part is exactly 0.5
For example try this
perhaps MrMath can shed some light on the subject?
not sure it’s a bug, it seems to be standard practice to round-off if the fractional part is exactly 0.5
For example try this
Code: Select all
a.double
b.double
F64_Val(a, "621.675001")
F64_Round(a, 2, b)
Debug F64_Str(b)
Posted: Thu Sep 08, 2005 2:48 pm
Could, perhaps, there be a roundUp() and a roundDown() function. Ceil() and Floor() or whatever?
Posted: Thu Sep 08, 2005 4:01 pm
There are already floor and ceiling functions but those will round up or down regardless of the fraction. I'm looking for a function that rounds down on a fraction like 1.1 to 1.4 while rounding up for 1.5 to 1.9.
@jack - Everything I've been taught since a kid is that you round up on 0.5 or greater. Excel does the same. "=ROUND(621.275,2)" will equal '621.28', rounding up on the 0.5.
@jack - Everything I've been taught since a kid is that you round up on 0.5 or greater. Excel does the same. "=ROUND(621.275,2)" will equal '621.28', rounding up on the 0.5.
Posted: Sun Sep 11, 2005 5:39 pm
Posted: Sun Sep 11, 2005 6:48 pm
Might if be an idea to make two functions?
One that treats 0.5 as a round up (as I was taught in school) ..
.. and one that treats it as a round down.
For some reason - and I cannot recall why now - we used to add 0.49999' to a value and truncate. So 0.5 when down and anything above went up. I may have been to overcome an error in early MS FP (which was non-standard way back when in days of DOS) or it may have been to do with some financial calcs.
Anyhow, reason for suggesting two is that if there are two ways, the lib will always be doing it the wrong way for someone, whichever way you go.
Then again, maybe I'm talking cr@p. I just have this feeling, though.
One that treats 0.5 as a round up (as I was taught in school) ..
.. and one that treats it as a round down.
For some reason - and I cannot recall why now - we used to add 0.49999' to a value and truncate. So 0.5 when down and anything above went up. I may have been to overcome an error in early MS FP (which was non-standard way back when in days of DOS) or it may have been to do with some financial calcs.
Anyhow, reason for suggesting two is that if there are two ways, the lib will always be doing it the wrong way for someone, whichever way you go.
Then again, maybe I'm talking cr@p. I just have this feeling, though.