RE: [boost] Submission policy

From: Beman Dawes (beman_at_[hidden])
Date: 2000-05-09 14:35:00

At 09:02 AM 5/8/00 -0700, Gary Powell wrote:

>Re: Portability,
>
>I'd hate to have a library rejected because a major vendor of a non
>compliant C++ compiler can't compile it. The hard part about
portability, is
>that its great in theory, and hard in practice, i.e. It's difficult
to spot
>the platform dependencies until you actually port the code.
>
>I'd like to claim that if a library can be compiled by two compilers
on two
>different platforms, that a basic level of portability has been
achieved.
>Even if I never get to use a library at least I'll have it to use as
a club
>with the compiler vendors, as in "Can your compiler compile the
boost
>library XYZ?"

Take a look at http://www.boost.org/more/lib_guide.htm  It supposedly
says more or less the same thing.

>As for acceptance into the standard, those two compilers and
platforms need
>to have a large audience so we can have a good trial of actual use
of the
>library.   Then the C++ standard library committee can decide
whether its
>worth adding the code to the standard.  After all there is a lot to
be said
>for having a small standard and a large body of useful
tools/libraries. Of
>course my libraries should be part of the standard because they are
great..
>:>

The standards committee (J16/WG21) marches to its own drumbeat, so it
is misleading to even speculate what if anything will move from Boost
to the next standard.  Boost members should consider joining the
committee if you really want to participate in the next standard.
Anyone who can (mostly) follow the technical discussion on boost has
the technical skills, and it is a really good way to build
professional experience.

>On the risk that my own submissions won't be accepted, I'd also like
to
>propose, that at least one person, possibly two need to nominate a
library,
>and those two people can't be the original authors. This way we can
see that
>at least someone other than the one resident expert recognizes the
value of
>a library. Face it, if you can't get at least two people interested
it
>probably doesn't belong in the standard.

That's part of the idea behind the concept of someone acting as
"review manager" (along with a need to spread the workload).

>I further propose that we have a links page, which we put links to
libraries
>which we like but don't want on the boost page. (i.e. too narrow in
scope,
>only work on single platforms, have restrictive licenses, need
perfect
>compilers etc.) That way more people will use boost as a first
source for
>their C++ library needs.

As long as it remains secondary to the main purposes of the site (the
boost libraries), several people have said they would like a links
page to other free libraries.  But I don't want to personally
maintain it.  If someone would volunteer to create and maintain a
links page, please speak up.  I think what most people have in mind
is a fairly small number of high quality links, so it shouldn't be a
huge effort to create and maintain.

--Beman