RE: [boost] Re: rationals initialization bug

From: Reggie Seagraves (reggie_at_[hidden])
Date: 2000-11-28 11:56:11

P.S.  The algorithm I gave you is an implementation of the
'law of continued fractions' developed by one of our engineers
in concert with personell from the Apple Numerics group.  We've
been using it in production code for some years now.  It's a
very 'subtle' algorithm.  I was given their tacit permission
to supply this algorithm to the boost project.

At 2:18 PM +0000 11/28/00, Moore, Paul wrote:
>From: David Abrahams [mailto:abrahams_at_[hidden]]
>> The checks you mention sound plausible, though. You know,
>> I'd be surprised if Knuth (or some other obvious source)
>> didn't contain a reliable version of this algorithm. It's
>> probably not a good idea for hacks like me (us?) to go
>> flailing about in numerics land without the theoretical
>> foundation ;-)
>
>I'll check my Knuth this evening, but I don't recall seeing this in there...
>
>I agree entirely about the theoretical foundation comment. My hidden design
>agenda for rational<> is to avoid including anything which requires a
>greater theoretical understanding than I have :-) [I'm a mathematician by
>training, a long time ago, but I never specialised in numerical maths...]
>
>Paul.