From: boost (boost_at_[hidden])
Date: 2001-05-28 04:57:13


Bonjour Hubert Holin,

On Saturday 26 May 2001 23:59, you wrote:
[..]

> Well, "catastrophe theory" is mostly concerned with the
> classification, stability and transformation of some kinds of
> singularites, but that's not very far.
And sqrt(x) is a simple, but important example.
>
> Indeed, there is some substantial corpus of knowledge in that
> field ("analytic" continuatioon, Riemann surfaces, etc.), even, if I
> recal, some dealing specifically with quaternions. It's just that I
> haven't had the time to correctly explore that aspect. And since I do
> that research in my free time while commuting (my daywork is hardly
> worth mentioning >-< )... and that I have a few other topics of active
> research... it may take some time before I can produce something
> worthwile in that field. But eventually I will tackle that...

You already have this problem with arctan<complex>.
I would suggest that it's sufficient to document the restirction on the
import. Once the's branch cut you just can't do it right.

Best wishes
Peter