Subject: Generalized Lambert W function - premature commit
From: David Billinghurst
Date: Thu, 24 May 2012 20:20:19 +1000
On 24/05/2012 5:47 AM, Raymond Toy wrote:
> Thanks for the updates. Everything works now.
>
> It's still a bit troubling that the bigfloat results aren't more
> consistent. Perhaps I'll try to track this down. It could be a bug
> somewhere.
>
>
> I have also used the bigfloat package to remove some duplicate
> code and there is scope for further improvement. I needed a
> bigfloat %e so I added the method to numeric.lisp, based on the
> code for the %pi method.
>
>
> Looks good.
>
> As an aside, the bigfloat package was intended so that you could use
> exactly the same code for floats and bigfloats, ignoring issues with
> constants and epsilon values. If you run into a case where this is
> not true, please let me know. (Of course, using the bigfloat package
> for floats will be slower than just coding the same algorithm for floats.)
>
> Ray
>
I have just checked in a generic version using the bigfloat package.
Much neater. I removed the guard digits from the bfloat calcs - it was
a nasty hack that didn't always work - so the results may change
slightly. I will just make a note in the manual that the final digits
may be wrong, and to suggest users check using higher precision if it
matters.