Re: [boost] [date_time] Who is in charge?

Subject: Re: [boost] [date_time] Who is in charge?
From: Robert Ramey (ramey_at_[hidden])
Date: 2012-06-30 14:35:24

Olaf van der Spek wrote:
> On Fri, Jun 29, 2012 at 12:02 PM, Klaim - Joël Lamotte
> <mjklaim_at_[hidden]> wrote:
>> I was surprised too, almost a year ago, that there have been no
>> reaction - not even a reply - to a date_time bug I reported on this
>> mailing list ( https://svn.boost.org/trac/boost/ticket/5753).
>
> Boost does not appear to have a policy for maintainers that go missing
> in action.
> I think someone (maybe steering committe) should come up with some
> policy.

well, I'm not on any steering commitee - as far as I know - but here's an 
idea.

let's take date/time as a good example.

a) Some smart guy with an interest/need for a version of the library with 
all
or most of the fixes applied does the work.

b) He adds his name to the copyright notice

c) Posts on his website and offers it to anyone who want's to pay
$X for a "supported/updated" version of the date/time library.  Payment
is via paypal or some simple system.

d) the boost version remains as it is, but now there's an alternative
for those who need an updated/maintained/supported version.

e) someday  - date/time INTERFACE becomes part of the C++ standard.
This means that compiler vendors who want to be "conforming" should
supply an implementation.  They now have a few choices:

i) roll their own
ii) roll in the boost version with minimal changes
iii) acquire rights to the supported/updated version

f) someday - the boost version falls into disuse as all/most compiler 
vendors
include a maintained/supported version with their products. The boost
version can now be dropped from the boost distribution.

Note that there is alternative date/time implementation - with a different
interface available from - I believe Rogue Wave.  So this isn't a totally
off the wall idea.

Everyone wins:

a) hackers like us can use the free version until it comes included with
the compilers we acquire.

b) companies which don't want to waste time/money fixing the package
have a cost/effective alternative.

c) someone gets paid to the the hard work.

d) C++ eventually gets a new standard library - compiler vendors have
a number of options from including an implementation along with their
product.

Any takers?

Robert Ramey