James Carlson wrote:

>[EMAIL PROTECTED] writes:
>
>>And I don't see how breaking up networking into further communities
>>helps achieve anything except to create more communities.
>>
>
>If they're more focused communities -- ones that are attempting to
>solve clear problems (such as "make the networking code go faster") --
>then they can steer projects in that area with greater effectiveness.
>
>Right now, we're without direction.
>

I fail to see how creating more, smaller, communities will solve
any of these problems.

The "without direction" problem requires leadership to solve.
And that's the problem - we have any number of "leaders" within
networking but no *leader*.  In addition, it isn't clear that any
of the "leaders" really has any responsibility inside of OpenSolaris
to do anything much.  ie. after pointing out that Solaris 10 is
mentioned on the networking community web page, nobody has
stepped up to fix that simple issue or even discuss how it should
be fixed - to that I'll add that when I have the Solaris networking
forum created on sun.com, I'll remove the relevant content from
opensolaris.org and add it to sun.com.

If smaller communities are constructed in the same manner as
networking is now, the real problems remain untouched.  We need
to examine the way we function, not just the groups.

Networking is unlike ZFS/SMF in that it is a composition of many
different aspects but each one of those is intimately related to
others.

If someone wanted to work on UFS in OpenSolaris, where would
they go today?  Should a new community be formed around that?
What about work on VFS?  Or does that suggest a new model
should be formed for filesystems and opensolaris communities?

The issue we have with networking is that there are many
different active fields, whereas in filesystems, the only really
active area is ZFS (and maybe CIFS.)

I suppose what I'm getting at is that I don't believe the
organisational structure of OpenSolaris into just "communities"
is sufficient.

At least one possible outcome that I think should be considered
is having SIGs (Special Interest Groups) inside communities.
I'd argue that there should be a filesystem community, within
which ZFS is a SIG (although that might offend ZFS people.)

And quite possibly another outcome is that communities are
required to elect or appoint leaders, whose roles and
responsibilities are yet to be defined.

Darren

_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to