Open jthompson6 opened 8 years ago
ln(1+ x)
is also preferable because it allows calculation of ln(1 - small positive number).
Group agrees to update the specification.
I made a pull request to fix it (although I don't know where else (other than Hadrian) that it needs to be updated.
This also needs to be updated on http://dmg.org/pfa/docs/library/; generate a new libfcns.html and put it on the website.
The spec for the function
m.ln1p
defines it asln(1+ x^2)
. This is surprising because the normal definition, isln(1+ x)
. This aligns with the classic Taylor series (the Mercator series) andjava.lang.Math.log1p
.Looking at the code for hadrian, I see that the implementation is actually
java.lang.Math.log1p
.I can verify this by running the PFA script
and seeing that the output for the first value matches the second, not the third.
So I think we should change the spec to say
ln(1+ x)
.N.B. Originally filed as https://github.com/opendatagroup/hadrian/issues/10