membership

["Roy T. Fielding" <fi...@liege.ICS.UCI.EDU>]

> Hmm. Out of curiosity, how do we define a "member of the Apache Group"? As
> far as I know, there's no ID card, or application form, or special
> handshake, or anything... If anyone asked me (which they haven't), I'd go
> for the IETF approach, which means that anyone on the list *is* a member.

Ummm, no -- the IETF does not allow anyone on the list to veto a patch
(change request).  We do, which means we need to at least believe that
the person has a clue and will be responsible *if* they vote.  That is
why we traditionally (but not always) invite someone to join after they
have supplied a few good patches and one existing member vouches for them. 
Mind you, this was a lot more visible when there were only nine people
able to do the inviting.

It used to be true that only people who were already invited to join
the Apache Group, or people invited to join in discussions (like NCSA
and RobM), were allowed to send mail to this mailing list.  Has that
been changed?  That is why the voting guidelines
<http://www.hyperreal.com/httpd/voting.html> say:

   Anyone on the mailing list may vote on any issue. However, the act of
   voting carries certain obligations -- voting members are not only
   stating their opinion, they are agreeing to help do the work of the
   Apache Project.

which has, in the past, been sufficient to differentiate between those
who vote and those who discuss.  However, it assumes that people are only
being added to the post-capable list after they have been invited to
join the group.  People who are invited to join should be informed of
the voting guidelines when they do.

Voting is a little more than just being a member of the group -- it also
means that you are willing and able to do the work.  That is why I don't
vote on releases except on those occasions when I have a working test
server and the time to do some testing.  Most of the Apache work that I
do is related to keeping an eye on standards issues, the W3C, and
occasional herding of the cats.  That, combined with reading all the
mail on this list and the web-related mailing lists, is why I vote on
anything related to standards or the future of the Apache project.
I am fairly certain that the rest of the group is following similar
guidelines when they vote, even though not all of them are written down.

For the lurkers out there, I'd say that it is safe to start voting when
you believe that the other group members won't be surprised when you do.
Vague as it is, that is probably the best way to proceed -- new members
have to convince existing members that their opinion is worth counting
as a vote (and, more importantly, as a veto) before they actually start
sending in +/- 1's.  When other people start counting your opinion as a
vote, even without any explicit vote in your messages, then that's a
good indication that it is safe for you to really vote within that topic area.

It actually isn't that hard.  Keep in mind that I've been around since
the beginning, but I've still only met Brian, RST, Dirk, and the NCSA
crowd in person (or at least that I recall) and I've probably done more
web-related conference travelling than anyone else in the group.  All the
rest I trust just on the basis of their work and e-mail communication.

If you don't know whether or not you should vote, just ask a few of the
old-timers -- if the response is "who are you?", then you shouldn't be
voting yet.

......Roy

p.s.  I expect the other group members to correct me if I got any of that
      wrong, or if someone holds differing expectations.

Re: membership

[Tony Sanders <sa...@bsdi.com>]

"Roy T. Fielding" writes:
> > Hmm. Out of curiosity, how do we define a "member of the Apache Group"? As
...
>    stating their opinion, they are agreeing to help do the work of the
>    Apache Project.
...
> If you don't know whether or not you should vote, just ask a few of the
> old-timers -- if the response is "who are you?", then you shouldn't be
> voting yet.
...
> p.s.  I expect the other group members to correct me if I got any of that
>       wrong, or if someone holds differing expectations.

As a non-voting member of the list how about this idea, you cannot
be a voting member of the list until you have submitted a patch
that was accepted or written a module or been a contributor to the
mailing list (for X amount of time?).  Or some variation on that
theme.  That is, someone who had actually ``help[ed] do the work
of the Apache Project'' in some form.

Re: membership

[Brian Behlendorf <br...@organic.com>]

On Thu, 27 Jun 1996, Paul Richards wrote:
> The list on the web site is very out of date, some people are basically
> not involved any more and some current contributors aren't listed. If
> we're going to have such lists then do it the FreeBSD way (check the
> web site). 

For the 1.1 release: if anyone out there is not listed on the contributors
page and would like to be, please send me mail with the appropriate
information to stick in place.

	Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com  www.apache.org  hyperreal.com  http://www.organic.com/JOBS

membership

[Paul Richards <p....@elsevier.co.uk>]

>>>>> ""Roy" == "Roy T Fielding" <fi...@liege.ICS.UCI.EDU> writes:

"Roy> It used to be true that only people who were already invited to
"Roy> join the Apache Group, or people invited to join in discussions
"Roy> (like NCSA and RobM), were allowed to send mail to this mailing
"Roy> list.  Has that been changed?  That is why the voting guidelines
"Roy> <http://www.hyperreal.com/httpd/voting.html> say:

"Roy> Voting is a little more than just being a member of the group --
"Roy> it also means that you are willing and able to do the work.
"Roy> That is why I don't vote on releases except on those occasions
"Roy> when I have a working test server and the time to do some
"Roy> testing.  Most of the Apache work that I do is related to
"Roy> keeping an eye on standards issues, the W3C, and occasional
"Roy> herding of the cats.  That, combined with reading all the mail
"Roy> on this list and the web-related mailing lists, is why I vote on
"Roy> anything related to standards or the future of the Apache
"Roy> project.  I am fairly certain that the rest of the group is
"Roy> following similar guidelines when they vote, even though not all
"Roy> of them are written down.

At some point things are going to get too big. I'm not going to push
the project one way or another but having been down this road before
eventually it's going to be necessary to create some "executive" group
along the line of FreeBSD/NetBSD/XFree86 etc. It's easy for a small
group of 10-15 to reach concensus in a finite time but when there's
100+ involved in the project things just never converge in a realistic
time period and decisions have to be delegated to a smaller body. Even
if this was to happen, it's extremely important to reach concensus with
the group as a whole or you start to piss off your contributors (having
a lot of contributors is generally a good thing). This is all totally
unrelated to the question of Apache as a legal entity, it's just a
matter of how the project is organised.

As far as coding decisions go, there already is such a body since to
actually get a change accepted into the src tree you have to convince
one of the committers that it's a good patch. I think that group of
people is currently very small and doesn't include people like Rob H so
it's not really a good forum for such decisions.

The list on the web site is very out of date, some people are basically
not involved any more and some current contributors aren't listed. If
we're going to have such lists then do it the FreeBSD way (check the
web site). 

The core team are listed as such, the development team (those with commit
priviliges) are listed as such (two relatively small groups), then anyone
who has ever contributed anything to the project is listed as a
contributor. It's not very fair to do things the way they currently are
on the "contributors" web page.

The way core works in the other projects I've been involved in, it's generally
been formed by the people who set the project up and are pushing the
project forward. As the project evolves new members are invited into core
by the existing core members (i.e. the core team makes a decision to bring
in someone new) usually because that person is working as hard as a core
member and therefore justifies their place.

Re: membership

[Brian Behlendorf <br...@organic.com>]

Thanks, Roy, for that excellent post.

On Wed, 26 Jun 1996, Roy T. Fielding wrote:
> > Hmm. Out of curiosity, how do we define a "member of the Apache Group"? As
> > far as I know, there's no ID card, or application form, or special
> > handshake, or anything... If anyone asked me (which they haven't), I'd go
> > for the IETF approach, which means that anyone on the list *is* a member.
> 
> Ummm, no -- the IETF does not allow anyone on the list to veto a patch
> (change request).  

I think he was referring to the goal, at least, of the IETF process, which
is to attain consensus, rather than having a straight vote.  If I were to
stand up during an IETF meeting and declare that I had a serious problem
with a proposal, it wouldn't really matter who I was or where I came from,
I'd still have the ability to speak, and then live with the consequences
if I don't make a convincing argument of course.  It's the only way a
couple of us college-dropout yahoos would have had any chance of
contributing to the process :)  But that is distinctly different than our
vote/veto system.

To me, the only distinction that exists from a functional level is between
those with write access to the RCS tree and those without.  While I have
been somewhat loose in telling people about the development group -
basically, if they look like they are going to be doing serious work with
Apache, then i figure the volume of the mail is enough of a filter - I
take seriously the granting of write access.  I trust those with RCS
access to be intelligent about the changes they make, to bring proposed
changes to the group first (and wait for feedback), and most importantly
to know how to rationally resolve conflicting desires.  This is all within
the context of a system which lets us back out of anything if there's ever
problems.  The write-access group is kept small to make sure things stay
manageable, but it's important to make sure there is enough activity from
those with write access that progress is indeed made. 

"Apache Group membership" is a term without any legal meaning right now.
I consider anyone who is a regular contributor to be a "member", if they
want to claim to be one.  Do newbies have a right to vote, to veto?  Sure,
but they have to be willing to stand up and defend their decision.  I
don't think anyone's ever vetoed something here withou giving an
explanation - we've had debates which could be called "religous" but
fortunately there's been enough humility to go around when it came to
implemention.  Hopefully that'll continue.

We can continue the discussion about the legal status of "the Apache
Group" after 1.1 is out - there are some hairy issues and politics 
involved, which I have been doing a fair amount of thinking about and
believe I can address, but now is not the time.

	Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com  www.apache.org  hyperreal.com  http://www.organic.com/JOBS