hypotenuse (was: Re: Adding a generic mathematical library)

[Christopher Dimech <dimech@gmx.com>]

> Sent: Monday, July 22, 2024 at 3:42 AM
> From: "Eli Zaretskii" <eliz@gnu.org>
> To: "Christopher Dimech" <dimech@gmx.com>
> Cc: incal@dataswamp.org, emacs-devel@gnu.org
> Subject: Re: hypotenuse (was: Re: Adding a generic mathematical library)
>
> > From: Christopher Dimech <dimech@gmx.com>
> > Cc: incal@dataswamp.org, emacs-devel@gnu.org
> > Date: Sun, 21 Jul 2024 17:12:10 +0200
> > Sensitivity: Normal
> >
> > > Sent: Monday, July 22, 2024 at 3:00 AM
> > > From: "Eli Zaretskii" <eliz@gnu.org>
> > > To: incal@dataswamp.org
> > > Cc: emacs-devel@gnu.org
> > > Subject: Re: hypotenuse (was: Re: Adding a generic mathematical library)
> > >
> > > > Date: Sun, 21 Jul 2024 17:44:47 +0300
> > > > From: Eli Zaretskii <eliz@gnu.org>
> > > > Cc: emacs-devel@gnu.org
> > > >
> > > > > (defun hypotenuse (c1 c2)
> > > > >   (sqrt (+ (* c1 c1) (* c2 c2))) )
> > > >
> > > > This is not the best way of computing hypot: squaring a number could
> > > > overflow or underflow, and you get bad result.  Try
> > > >
> > > >     (hypotenuse 1.e-600 1.e-600)
> > > > or
> > > >     (hypotenuse 1.e600 1.e600)
> > > >
> > > > and you will see it.
> > >
> > > Sorry, that should have been 1.e-300 and 1.e300 instead.
> >
> > I have a solution to the problem in fortran, but not in elisp.  I could
> > include it.  Its implementation required serious concentration.
>
> AFAIK, every libm already includes a similar implementation, so we
> should just expose it to Lisp (again, assuming this function is indeed
> required to be part of Emacs).

The people involved in calc do think it is required.