"Peter v. N." <sk*********@hotmail.comwrites:
This may sound heretical: but I'd like to know what the community
thinks about using Boost extensively in C++ projects (mainly for
threads and numeric applications).
My company does aerospace engineering and algorithm research; we
use Boost extensively. Our code is extraordinarily numeric, and
we're moving towards threads.
The only reason not to use Boost that I can possibly imagine is for
flight-grade embedded code, just because Boost hasn't been certified
at that level (to my knowledge). We're moving towards developing
flight grade code in the next few years, and that will be an issue.
Aside from specific situations like that, IMNSHO any C++ programmer
who doesn't use Boost is crazy.
Go look at the "Who's Using Boost" link on the Boost website.
Better reinvent the wheel (for true C++ programmers) or depend on
third party frameworks?
True C++ programmers _never_ reinvent the wheel.
True C++ programmers inherit publicly if my_wheel IS-A wheel, and
they aggregate if my_tire HAS-A wheel, and they refactor when it
becomes apparent that all wheels should have had a
my_momentOfInertia to begin with.
True C++ programmers may rebuild the wheel, if the wheel they are
provided isn't robust enough. Boost is just about as robust as
libraries come.
Good libraries, using OO | functional | generic | meta- programming
when appropriate... that's what C++ was designed for.
I know that with the upcoming release of C++ many parts of Boost
will be incorporated...Nevertheless I'm curious about what you
think.
I think that there are good reasons Boost is pushing the C++
standard, mainly that some people on the Standard C++ Library group
started Boost for exactly this purpose.
----------------------------------------------------------------------
Dave Steffen, Ph.D. Disobey this command!
Software Engineer IV - Douglas Hofstadter
Numerica Corporation
dg@steffen a@t numerica d@ot us (remove @'s to email me)