Sam Ruby wrote:
Suppose somebody new contributes 1K lines of quality code a week to Maven over the course of a year... and is voted in as a comitter. Clearly a CLA is required, but does a PPMC need to be created or does the incubator need to be involved?
Of course not, as you well know. He is contributing to an existing ASF codebase. I assume that this is rhetorical to setup your argument. :-)
Guilty. But in my defense, it is my experience that posing a specific thought experiment grounds discussions sufficiently to enable discussions to make forward progress with a minimum of theoretical tangents on abstract concerns. Let's see if it works:
Suppose I write 50K lines of code and it lives on sourceforge for a year. Maven takes a liking to it and I contribute it (and in the process am made a committer). Clearly a software grant is required, but is a PPMC needed or does the incubator need to be involved?
As I understand the Board's creation of the Incubator, yes. Any incoming codebase must go through the Incubator. I do think that some projects may be spending too much time in the Incubator, though. Not pointing figuers, just wondering why some of them haven't completed the necessary steps to leave.
And, I, in turn, think that some have interpreted the Incubator's mandate too broadly. Not pointing fingers or anything :-P
Actually I do think that the PPMC proposal, once defined and refined, will address my concerns nicely.
Mix scenarios 1&2. Somebody new to the ASF writes the code on sourceforge, wishes to donate it to an existing and vibrant community, signs both a CLA and a code grant. PPMC? Incubator?
Is this a new project, or a large set of patches for an existing one? If it is a new sub-project, it has to go through the Incubator. If it is just a set of patches to be applied into an existing codebase, I don't think that you could bring it "into" the incubator, but we still have an issue in terms of making sure that we have clear title to the code.
No question on having the incubator "vette" the clear title. I actively support that role for the incubator on all contributions.
My presumption is that existing, healthy communities can be presumed to make incremental increases in the number of committers without requiring the incubator to be involved
Agreed.
But I *would* like the incubator to be involved with all code grants and with the formation of all new PMCs (by that I mean truly new, not merely ones formed by bifircation).
Does this make sense?
Yes. The only point of difference appears to be the importing of sub-projects. As I understand it, the Board has said that the existing PMCs are not to self-incubate incoming codebases. And we should agree that the Incubator would be happy to release projects as soon as they are ready.
Can you provide a reference? Remember, I was there. ;-)
I still believe that there is a semantic gap between how I am using the term project and how you are using the term project.
The board does not formally recognize the existance of subprojects in any way. Nor does it meddle with project affairs at that level. What the board does require is that projects (i.e., those things listed in cvs:committers/board/committee-info.txt) be able to demonstrate active oversight of the code that they are responsible.
- Sam Ruby
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
