[boost] Re: Adding Loki to Boost (reprise)

From: Andrei Alexandrescu (andrewalex_at_[hidden])
Date: 2002-04-14 20:55:18

"Aleksey Gurtovoy" <agurtovoy_at_[hidden]> wrote in message
news:a9d7rv$t7o$1_at_main.gmane.org...
> How about this one:
>
> template <class L, class R>
> struct InheritTwo : public L, public R
> {
> };
>
> template< typename Sequence, typename F >
> struct GenScatterHierarchy
>     : mpl::fold< Sequence, mpl::none, InheritTwo< mpl::apply<F,_1>,
_2 >
> {
> };

The versions of GenScatterHierarchy and count_if that you posted
lately look better indeed. They do rely on quite some mpl artifacts -
much more than I find palatable for what they do -, but I guess
there's no reason to agonize over that.

> > This being said, an implementation of GenScatterHierarchy that
works
> > only for dot-typelists and relies on nothing but C++ itself, has
20
> > lines, defines 1 new type, and doesn't use but the typelist
itself.
>
> FWIW, the MPL's version above is 9 lines long. Much more
importantly, it's
> something that is explainable to a non-expert in 5 minutes. (IMO
Loki's
> version is not).

(In the meantime I realized I could have reduced the length of
GenScatterHierarchy by eliminating the unnecessary LeftBase and
RightBase convenience typedefs. But oh well.)

I believe we reduced irreductible viewpoints about the ease of
explanation of the two idioms, so I won't insist on it anymore. The
truth is simple: we are both right. At least because for people like
you, one version is more appealing, and for people like me, the other
is.

Andrei