iSOblog

i tried to better understand how float values are stored on a computer so i decided to write a little class that should help me with this.

IMPORTANT NOTE: i can’t be held responsible for anything that this class does or does not, you’ve been warned :p

list of know issues:

• the methods floatUnderstandingFromLongValue() and floatUnderstandingFromDoubleValue() return WRONG FLOATS!
• maybe the method getMantissaBitsSignificanceString() is also not working correctly, especially with really small numbers (E=-126)
• code has gotten a little confusing
• things could probably be coded more efficiently

you can download the source from here:
http://pascal-schwarz.ch/stuff/FloatUnderstanding/FloatUnderstanding.java
http://pascal-schwarz.ch/stuff/FloatUnderstanding/TestRecalcMultiThread.java

if you have any feedback, please use the comment-functionality under this post.

Update (20.03.2011):

• better handling of denormalized values (getMantissaValue(), getMantissaSignificanceString(), getExponentValue(), …..)
• included a multithreaded recalcFloatValue-Tester
• the fabric-methods floatUnderstandingFromLongValue() and floatUnderstandingFromDoubleValue() ARE NOT FIXED yet