From: Paul A Bristow (pbristow_at_[hidden])
Date: 2007-12-13 04:34:12
>[mailto:boost-bounces_at_[hidden]] On Behalf Of Hugo Duncan
>Sent: 12 December 2007 23:45
>Subject: Re: [boost] [rfc] Quadrature library for numerical
>On Wed, 12 Dec 2007 Paul A Bristow wrote:
>> Did you notice any improvement in compile time by using argument
>Not as much as I expected, though that may be due to how I have
>implemented the chaining - each time an argument is set, a new
>object is returned (with a slightly template instantiation).
>> PS Warning: C++ Purists may choke and splutter at finding
>> code in an .hpp file, but it's essential documentation from
>> the 'bottomless pit of working subroutines'.
>Oops, did I leave some in. As far as I remember, only in the
>integrals.hpp test file.
No, I was joking, but many a true word spoke in jest -
it IS a vital part of the documentation.
(though perhaps it need not be part of the Boost *distribution*,
which reaises an interesting question about where background material should live. It needs to be somewhere *permanent* - which
perhaps is not the intention or reality for the Boost Vault?)
>> PPS Of course it would even nicer with a templated RealType
>- but don't panic - version 2...
>It should already work with a templated real type. I am
>trying to get it
>to work with Boost.Interval, and have downloaded NTL to have
>go with that as well.
You will find essential John Maddock's patches for NTL used with the Math Toolkit - in the sandbox until 1.35 is out. And it would
be really nice if we could persuade the author Victor Shoup to release at least the base part of NTL under the Boost license - is
GPL at present.
--- Paul A Bristow Prizet Farmhouse, Kendal, Cumbria UK LA8 8AB +44 1539561830 & SMS, Mobile +44 7714 330204 & SMS pbristow_at_[hidden]
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk