From: Dan W. (danw_at_[hidden])
Date: 2004-01-09 16:21:51
Matthias Schabel wrote:
> As an aside, what's the interest level in having dimensions be
> specified by
> arbitrary arithmetic types (rather than static_rational as I currently
> do)? I'm
> pretty sure I'll be able to remove the dependence on static_rational in
> the
> dimensional analysis code so that any class for which these compile-time
> arithmetic operations is defined would work :
>
> static_negate<T>
> static_add<T,T>
> static_subtract<T,T>
> static_multiply<T,T>
> static_divide<T,T>
> is_zero<T>
I think it would be ultra-wise to have such de-coupling, even if there
were no immediate interest at boost developers forum. The fact is, the
more forward-looking the design is, the more promise it holds.