On Wed, 2002-11-13 at 04:20, Rodent of Unusual Size wrote:
not strictly true, although mostly. a product release may be effectively vetoed by the asf officer with oversight of the project, if it appears in that person's judgement that releasing it would be the Wrong Thing
I'm curious - are those rules invented as we go ?
Sam - you are the PMC chair for jakarta, did you know that rule ? Or the rules about direct ( and non-delegating ) oversight, or the non-protection for non-PMC members?
You ask several questions here. Let me answer them one at a time.
In http://jakarta.apache.org/site/pmc/01-01-17-meeting-minutes.html, a meeting that you were in physical attendance, one of the roles of the PMC was to act as the veto of last resort. One such instance where I would personally exercise such a veto is if the majority of committers to Tomcat were to want to create a release which is not compliant with the servlet APIs. This is not the only reference I have made to the phrase "veto of last resort".
I am aware that the role defined for the chairman of a PMC is a special one. One that I do not, as a rule, chose to draw attention to.
To be perfectly honest, I am not clear on rules which outright preclude delegation. However, one thing I will observe is that Jakarta has not effectively delegated the oversight of Avalon, something that must be corrected ASAP.
As to the non-protection of non-PMC members, IANAL. But one thing I am pretty sure of is that if we cannot establish a clear set of accountability, the ASF itself is exposed.
I would feel much better if you did - and decided to not tell the rest of us for whatever reasons. Because if you didn't - I think ASF has a very serious problem.
I do believe that the ASF has a number of challenges at the moment. And in my perception of the scheme of things, the topic referenced by this subject line doesn't quite rank.
Nearly two years ago, there was a near-complete fork of Tomcat. In consultation with Brian and Roy, the Jakarta PMC worked with the developers of Tomcat to build a plan. That plan explicitly allowed further releases of Tomcat 3, even with new functionallity. The hope was that ultimately the result would be a converged Tomcat 5. Now that that goal is within our collective grasp, there seem to be some who wish to reinforce the rapidly dampening waves. Ultimately, they will be unsuccessful.
What is a more significant challenge that the ASF need to address is the problem of accountability and effective oversight. The proper path (again IMHO) is to establish groups that can be trusted to be self-governing and self-regulating. In my admittedly biased opinion, Jakarta largely meets this criteria, with some some glaringly obvious exceptions that are being addressed.
- Sam Ruby
