geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Aaron Bannert <aa...@clove.org>
Subject Re: Geronimo PMC
Date Fri, 12 Sep 2003 18:44:33 GMT

On Friday, September 12, 2003, at 11:22  AM, Jeremy Boynes wrote:

> The discussion on the process for nominating committers has highlighted
> confusion about the role played by the PMC for a project and I believe
> before we can make process on procedures for Geronimo this must be
> clarified. This may be stepping into a political and social minefield,
> but events over the last week have shown this is something that must be
> resolved.
>
> There appear to be two different approaches within the ASF to the role
> played by the PMC:
>
> Firstly, there is what, for want of a better name, can be called the
> httpd approach. With this structure, the PMC has a relatively large
> number of people comprising the active committers on the project.
> Technical decisions are made by the committers on the public 
> development
> list, but procedural decisions (such as adding a new committer) are 
> made
> in private on the pmc list.
>
> Secondly, and again for lack of a better name, there is the Jakarta
> approach. Here, the PMC is smaller and seems to deal mainly with
> organizational issues such as adding projects, ensuring CLAs are filed
> and co-ordination between sub-projects. Both technical and procedural
> decisions are made in public on the sub-project development lists and
> then passed to the Jakarta PMC where appropriate.
>
> I would speculate some of the confusion arises because many members of
> the Geronimo community are new to Apache and, being Java centric, are
> more used to the Jakarta approach, whereas the incubator PMC is more
> used to the httpd approach.
>
> My first question is whether this is a fair and accurate summary? If I
> am just confused, please just ignore the rest of this mail.

This is a very good characterization. +1 :)

> The second one is whether the ASF as a whole has a formal preference 
> for
> one of these two approaches, or has prior experience shown one to be
> "better" (with allowance that there probably as many opinions on that 
> as
> there are community members)?

I believe the ASF as a whole would like to see new projects adapt
the httpd/apr method of private PMC nominations and discussions of
new committers and PMC members.

[entering minefield]
The incubator was created in part to avoid the problems of oversight
that the Jakarta PMC was having over its subprojects. By encouraging
contributors to become committers, and committers to become PMC members,
the "community" becomes spread over the spectrum of responsibility based
on merit, rather than concentrating the responsibility for oversight
in a few overworked individuals. By spreading around the responsibility
(based on merit) we also allow the ASF to scale and add more and more
subprojects and top-level-projects that are self-managing and 
successful.

> The third question is whether either of these is the "right" solution
> for Geronimo, and if not, then we need to define an organization
> structure that meets the legal and philosophical requirements of ASF 
> and
> the needs of the Geronimo community.

Well said! +1

> The Jakarta sub-project concept seems to fit with the current status of
> Geronimo - it is not yet a project in its own right, but a sub-project
> of incubator. Given that, it seems to make sense to have the project
> operate in a similar manner, where the committers make the technical 
> and
> procedural decisions and, where necessary or appropriate, pass the
> results to the PMC to execute.
>
> However, we still need to debate and define the management structure
> that will be put in place when Geronimo leaves the incubator. Either of
> the two approaches may be appropriate, or we may need a different
> variant capable of handling the legal requirements imposed by
> certification.
>
> This discussion should start now but we cannot place Geronimo on hold
> whilst it is resolved. There was very strong support in the vote on the
> committer process to "do it a standard way", and the way chosen was 
> that
> of a Jakarta sub-project. Is it the right way, I think not - is it good
> enough for now, I believe so.

Unfortunately, there is no standard new committer process. We can only 
adapt
(and improve where necessary) what has worked well in the past.

-aaron


Mime
View raw message