[ https://issues.apache.org/jira/browse/MATH1386?page=com.atlassian.jira.plugin.system.issuetabpanels:alltabpanel
]
Gilles updated MATH1386:

Priority: Trivial (was: Major)
Issue Type: Wish (was: Bug)
> EPSILON value in org.apache.commons.math.util.MathUtils seems like half of what it should
be
> 
>
> Key: MATH1386
> URL: https://issues.apache.org/jira/browse/MATH1386
> Project: Commons Math
> Issue Type: Wish
> Environment: win 7
> Reporter: Brian S. McCormick
> Priority: Trivial
> Original Estimate: 1h
> Remaining Estimate: 1h
>
> I've always used 1.1920929E7 for dealing with float epsilon values. When dealing with
double I've always used 2.220446049250313E16 for epsilon.
> You have it defined in org.apache.commons.math.util.MathUtils as 1.1102230246251565E16
which is half what I think it should be.
> I come up with these numbers using the following:
> float fEps = Float.intBitsToFloat(Float.floatToIntBits(1f) + 1)  1;
> double dEps = Double.longBitsToDouble(Double.doubleToLongBits(1) + 1)  1;
> Am I correct? I don't really know. I do know that float epsilon in every legacy C/C++
compiler etc I have ever used is defined as about 1e7 and this is the value using the formula
for fEps above. When I started doing comps using doubles instead of floats I started using
the formula for dEps above which looks to me like the equivalent for double numbers.

This message was sent by Atlassian JIRA
(v6.3.4#6332)
