* John Plocher <[EMAIL PROTECTED]> [2005-08-10 09:39]:
> Roy T. Fielding wrote:
> >At some point we need to make a distinction between creating
> >mailing lists/forums and creating self-governing groups responsible
> >for building products.  Which one falls under the name "community"?
> 
> +1!!!!
> 
> IMHO, IAC!
> 
> I thought I had a handle on this community thing when we launched the
> OpenSolaris "ON Community" - a group of people involved with the
> "source tree formerly known as the Solaris Kernel Consolidation".
> 
> The proto-governance proposals floating around all made sense to me
> when used in that context, though it was easy to imagine other equally
> as useful interpretations.  My "sanity test" was whether the following
> phrase rang true:
> 
>       The community consists of the movers and shakers (aka maintainers)
>       who are steering, architecting, designing, implementing and 
>       sustaining
>       the codebase in question.
> 
> This implies (to me) that a community is something bigger than a single
> bugfix, yet smaller than an entire distro.  My experience with things at
> Sun leads me to associate "a community" with the abstraction we call a
> "consolidation", and not with smaller "projects" or larger "products".
> 
> NB: The jury is still out as to whether or not this association is valid.
> It could be that the thing we currently have is "too big" to be a single
> community, and needs to be refactored into smaller abstractions.  Some
> possibilities that come to mind are splitting out drivers, filesystems,
> posix/gnu utils and the core kernel/networking stacks - each would have
> its own community driving its growth, independent of the others.  Or not.
> 
> Then came the explosion of mailing lists and forums on the web site.  They
> were also called communities, yet not all were focused on things that were
> directly tied into "identifiable pieces of code".

  I had been reading the governance draft such that communities had
  a precise membership and one or more formal ways of achieving
  consensus.  Like John, I had matched up communities with a number of
  the bodies in the internal process:  some communities were
  consolidations, some were more like steering committees ("doing X is
  important"), and some were technical programs.

  It does seem like there is a need to provide some mechanism supporting
  group collaboration that is more ad hoc than the community.  Let's
  call this a "project".  A project could have a mailing list and a home
  page and, eventually, a source code repository of some kind.  Project
  initiation and membership do not necessarily connect with any of the
  communities, but a community could choose to endorse one or more
  projects (which would show projects with some consensus support).

  For instance, I would expect smf would be turned into a project, and
  that it would gain the endorsement of the Nevada and (I hope) the
  Approachability communities.  

  If opensolaris offered projects with mailing lists (and we revisited
  some of the current communities and converted them to projects), would
  that support the specialized conversations people want to have, while
  preserving the communities as the decision-making bodies over larger
  products and aspects?
  
  - Stephen

-- 
Stephen Hahn, PhD  Solaris Kernel Development, Sun Microsystems
[EMAIL PROTECTED]  http://blogs.sun.com/sch/
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org

Reply via email to