Re: firls.m, part 2

[Mike Miller]

On Mon, Jun 19, 2017 at 16:23:45 +0000, je suis wrote:
> > I suspect that it might, depending on what you mean by "adapted from".
> > You may want to take down your code wherever it is hosted publicly. Take
> > a look at http://numerical.recipes/licenses/redistribute.html.
> 
> I see, but the underlying mathematics is based on Abramowitx & Stegun.
> If I modify the code, it should be safe then, shouldn't it?

There is no right answer and I am not a lawyer. I also haven't read NR,
I haven't looked at the relevant A&S section, and I haven't looked at
your code. The expint function in Octave has comments indicating parts
of it are also based on A&S. My recommendation would be to start from
the current expint function and make the necessary changes to it to fix
the problem you have found, based on what you have learned about the
algorithm.

-- 
mike