Re: hal++ announcement

[Marko Anastasov <marko marko anastasov name>]

On Sun, 2006-11-26 at 22:30 +0200, Pavlo Korzhyk wrote:
<snip>
> As for signals - I would say "one lib to rule them all". Using Boost
> must be as natural as STL, and it's signals are really well designed &
> cool. For me it's a disadvantage of gtkmm and the rest that they
> something else instead of boost.signals.
> We must develop new libs, implement _new_ ideas and not waste time,
> resources and end-users learning time on different implementations of
> the same idea.

Personally (and I believe I'm not alone) I find using sigc++ very
natural too. IIRC about a year ago boost (?) people asked the sigc++
developers to start working on what will become a signalling framework
in the future version of STL. My point is that both are not far from
the future standard.

In my opinion, "one lib to rule them all" in our world actually already 
applies to sigc++, and it wouldn't be a very good idea, as Milosz has
said, to force boost as a dependancy. The long term goal for hal++
should be to become the official C++ bindings for HAL, and somewhere in
between someone will maybe even start wrapping D-Bus (as it's passed 1.0
now). Basically, the GNOME C++ bindings use sigc++, and all the related
new ones should follow, until a new C++ standard comes alive.

Marko