On Feb 8, 2007, at 2:43 PM, Ceki Gülcü wrote:
If I understand correctly, the idea is to circumvent the quorum
requirement, i.e. collecting three +1 votes, so that the Logging TLP
can vote instead of the people who are actually familiar with the
work?
From http://www.apache.org/foundation/voting.html
Binding Votes:
Who is permitted to vote is, to some extent, a community-specific
thing. However, the basic rule is that only PMC members have
binding votes, and all others are either discouraged from voting
(to keep the noise down) or else have their votes considered of an
indicative or advisory nature only.
That's the general rule. In actual fact, things tend to be a little
looser, and procedural votes from developers and committers are
sometimes considered binding if the voter has acquired enough merit
and respect in the community. Only votes by PMC members are
considered binding on code-modification issues, however.
From http://www.apache.org/foundation/how-it-works.html#structure
The role of the PMC from a Foundation perspective is oversight. The
main role of the PMC is not code and not coding - but to ensure
that all legal issues are addressed, that procedure is followed,
and that each and every release is the product of the community as
a whole. That is key to our litigation protection mechanisms.
From http://www.apache.org/dev/release.html
Releases are, by definition, anything that is published beyond the
group that owns it. In our case, that means any publication outside
the group of people on the product dev list. If the general public
is being instructed to download a package, then that package has
been released. Each PMC must obey the ASF requirements on approving
any release. How you label the package is a secondary issue,
described below.
Each of those makes the PMC responsible for releases. The PMC must
vote on a release and while it definitely should take in to
consideration any vote or opinion by non-PMC committers or community
members, it can't delegate responsibility for ensuring that the
release follows ASF requirements and licensing guidelines. If a PMC
member is not familiar with the work and have not checked it for
sanity and legal issues, they should not vote on a release.
Requiring a two-stage vote (once at the subproject level and once at
the PMC level), either adds an almost identical vote (in log4j's case
where the PMC and log4j committers are very close to the same set of
people) or delays general PMC review of a release until 72 hours
after the build has been proposed on a subproject where only a few
PMC members are active. Either situation can be addressed by either
calling the PMC and project votes simultaneously or by PMC members
joining a subproject vote, however those are ugly, but effective,
hacks. The two-stage vote in my opinion adds complexity without a
corresponding benefit.
For chainsaw, it has resulted in a weird
hybrid where chainsaw has its own project in the SVN and releases,
but is part of the log4j project for quorum issues (and potentially
other issues).
That was by volition of Scott and Paul. At the time, I invited
them to form a separate project and they declined. You should perhaps
ask Scott and Paul whether they would like to become a separate
sub-project.
Hopefully they are subscribed to this list and will participate in
the thread.
In addition, there are some definitions in the bylaws
(IIRC some of the voting definitions) that conflict with other
definitions in more authoritative Apache documents.
There ASF is based on delegation. Therefore, as long as the LS bylaws
are in agreement with rules of the ASF, they are authoritative as far
as the LS TLP is concerned.
The Foundation glossary defines "Lazy Approval" and "Lazy Consensus"
as synonyms while the LS Bylaws has them as distinct terms with "Lazy
Consensus" with a similar definition as "Consensus Approval".
http://www.apache.org/foundation/glossary.html#ConsensusApproval
http://www.apache.org/foundation/glossary.html#LazyConsensus
If the bylaws are modified, it would be good to change occurrences of
"Lazy Consensus" to "Consensus Approval" to be consistent with other
ASF projects.
There have been previous discussions on bylaw changes to address
these issues, for example on [email protected] on 2005-07-01
(http://marc.theaimsgroup.com/?l=apache-logging-
general&m=112024592311861&w=2) which referenced earlier discussion on
the private pmc list, but there was no action taken.
The Board Resolution forming the LS PMC mandated the initial PMC
create a set of bylaws for the project (http://logging.apache.org/
site/mission-statement.html). The term "guidelines" appears to be
currently preferred for project level statements to avoid the legal
implications of bylaws. For example, the ASF Bylaws (http://
www.apache.org/foundation/bylaws.html) are a legal document that
governs the corporation and has to conform to corporate governance
laws in the State of Delaware.
The LS bylaws (see http://logging.apache.org/site/bylaws.html )
mentions the word "guideline" once.
The point was that the term "bylaws" has specific legal connotations
and recent board discussions had preferred using the term
"guidelines" for PMC level documents and reserving the "bylaws" term
for the ASF bylaws which is a legal document.
Section 6.3 of the ASF Bylaws (http://www.apache.org/foundation/
bylaws.html) places responsibility to establish rules and procedures
on the PMC chair.
The foundation bylaws mandates the project chairman to establish
project bylaws. However, once the bylaws are established, they should
not and cannot be changed at the whim of one person. We already have
bylaws. They can be changed as long as there is a 2/3 majority of the
active PMC members. The 2/3 rule is there precisely to prevent the
chairman from making arbitrary changes.
Are you suggesting that the chairman has the power to change the LS
project bylawys without approval of the PMC?
This reference was provided for background material on project
governance. It was not a threat of unilateral action.
There was an discussion on [email protected] on 2005-09-12
that discussed the confusing state of ASF project bylaws (http://
thread.gmane.org/gmane.comp.apache.incubator.general/6144/
focus=6147).
I thought that I recently saw a thread on
[email protected] that argued against projects having
bylaws (or guidelines) of their own as the ASF bylaws and other
documents were sufficient, but unfortunately I have not been able to
find that thread.
As an open group, people are welcome to argue for or against any
topic. The existence of a con argument does not mean it should be
adopted here.
I was trying to enumerate all the pertinent discussions that could
inform our discussion. Unfortunately, I was unable to find a
reference, but thought a description of the thread might be enough
for someone else to fill in the details.
I would like to take steps that result in:
The LS project to be a single project with any number of products.
The PMC being the only decision making body.
Either no project guideline document or one that heavily defers to,
not duplicate or conflict with, http://www.apache.org documents.
A set of evolving process documents that describe best practices on
the project.
I would appreciate comments and will try to work on a draft. A board
report is due for the Feb 21st meeting. It would be great to get
this and several other issues resolved in the next week or so in
order to clear out some long pending issues.
The idea behind the current structure is to delegate responsibility to
the people actually doing the work. I still think that delegation is
one of the core tenets of the ASF. Thus, after a given LS sub-project
votes, and then makes a recommendation to the PMC, the PMC is very
likely to endorse the recommendation. As far as I know, the LS PMC has
always endorsed recommendations made by sub-projects.
Instead of undoing the existing bylaws, I would suggest to expedite
the PMC voting process. Moreover, nothing prevents a sub-project from
granting committer access to an existing PMC member (hence mitigating
the difficulty of reaching quorum.)
Thanks for your comments.
