[math] About MATH-849 and the NSWC library

[Sébastien Brisard <se...@m4x.org>]

Dear all,
please review the code I've attached to MATH-849. The improvement in
accuracy is great. I also implemented, besides logGamma(x), a new function
-- namely Gamma(x).
I would like to know if you feel the references to the NSWC library is fair.

Thanks for your advice,
Sébastien

PS : with this new impl, some tests now fail in GammaDistribution (as
expected, this is related to MATH-753). I'm on it.

Re: [math] About MATH-849 and the NSWC library

[Sébastien Brisard <se...@m4x.org>]

Hi Thomas,

thanks for reviewing.

>
> static final double HALF_LOG_2_PI is not used anymore
>
well spotted! I'll get rid of that. I'm also currently revisiting
MATH-753, I would like to get rid of all things related to the Lanczos
approximation (which is no longer used for Gamma, but still necessary
for GammaDistribution -- for the time being).

Sébastien

PS: since you've already had a look to the code, and in order to get a
broader review on the references to NSWC in the javadoc, I'm going to
commit the code soon, and update the CM website, so that I can point
to it directly in a thread. What do you think?


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org

Re: [math] About MATH-849 and the NSWC library

[Thomas Neidhart <th...@gmail.com>]

On 08/22/2012 06:43 AM, Sébastien Brisard wrote:
> Dear all,
> please review the code I've attached to MATH-849. The improvement in
> accuracy is great. I also implemented, besides logGamma(x), a new function
> -- namely Gamma(x).
> I would like to know if you feel the references to the NSWC library is fair.

Hi Sebastien,

I have looked briefly at the code so far, but apart from a very minor
thing (static final double HALF_LOG_2_PI is not used anymore), it looks
pretty good imho.

Great work!

Thomas

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org