Re: [boost] Boost.MathConstants: Review

From: boost (boost_at_[hidden])
Date: 2001-04-23 05:37:46

Salut

On Sunday 22 April 2001 13:01, John Maddock wrote:
[...]
>
> Some points I haven't seen mentioned yet:
>
> The docs provide a nice rationale, but don't indicate what constants are
> actually available (perhaps the most important point?).
>
> Defining the constants as members of a template, allows us to add
> specialisations for types other than built in floating point types - for
> example if we add boost::interval or boost::bcd, then these should have
> math constants available as well.  
I mentioned the interval classes earlier, but maybe it was not clear in 
enough. Once there is a interval-type availbale in C++ I would like to
have an enclosure of  the math constants, e.g.

const interval<double> boost::math::constants<interval<double> >::pi =
        intervall<double>( 3.1415, 3,1416);

>> const intervall<double> pi = (double)BOOST_PI; /* C style cast */"
Just makes no sense here.

> BTW what ever happened to the boost
> interval library?
Is there work on an interval library?
That would be really great.

> I prefer "boost::constants<T>" to the "boost::math::constants<T>" which
There are so many constants living in our world, even constants which are
not really constant (depeding on the the acuracy of measurement, even worse
"running constants"), so I'd prefer to have mathematical constants in
boost::math. THey should neve be redefined.
Constants that are measured in kilogramm, meters, seconds ... may
get new official values. Note that the definition of the kilogramm is not
very accurate compared to meter and seconds.
Those constants could live in boost::scientific.
In addition I have really paid a lot for accidental name clashes,
so I don't care type a few extra letters.

Best wishes,
Peter

P.S: In case you are not sure if pi is really a constant you should read 
"Fashionable Nonsense: Postmodern Intellectuals' Abuse of Science"
Alan Sokal and Jean Bricmont.