Re: Process for becoming a mentor

[Erik Abele]

On 23.03.2006, at 14:36, Jason van Zyl wrote:

Erik Abele wrote:

On 22.03.2006, at 18:01, Jason van Zyl wrote:

Hi,



1) Active interest in the project being mentored

Just poking around the incubator site and I see a responsibilities page for Mentors which is good but I don't see any criteria for being a mentor. But I think starting with three simple ones would be a start:

This makes obviously sense but...

2) Commitment to future involvement

3) Discussion with the project to see if there is a good fit and get the buy in of the folks on the project

...these two don't - furthermore I don't see any value in more and more policies and criteria in this case.

I respectfully disagree. I absolutely believe 2) is desirable if only as a matter of respect and courtesy but more importantly I think it can only make the incubation process easier.

Of course, nothing bad about that but not as a _requirement_ for being a mentor.

As per 3) if you look at the projects page I think you see mentors who are involved in the incubating project and plan to be involved past its incubation. The last project to graduate was JackRabbit which I believe will be successful because Roy was involved from the beginning, is active in the project and will continue to be I'm sure.

Well, it's his 'dayjob' ;-) but seriously these are *not* criteria or requirements for being a mentor. I'm so anal about that because there were quite a few people here saying: "java? I don't know java so I won't be of any help as a mentor." and that is simply not the correct approach here.

ADF Faces and Cayenne look like good examples of this as well.

Hmmm, although I think we should basically welcome any new mentors with open arms, the above description sounds more like a committer than a mentor. Are you really sure that you understand the tasks and responsibilities of a mentor?

I think so, I've been here for 5-6 years and helped projects like OJB, Lucene and XmlRpc come to Apache though this was prior to the formation of the Incubator.

Ok, but go back and read your message again: you didn't mention that you want to help with your experience and knowledge in community building - you were proposing technical changes/improvements only - that just doesn't fit with a mentor! If these are your goals [the 2) bullets of your msg] then you'll just need to subscribe to the dev@ lists, send in patches and get invited as a committer - you don't have to be a mentor for that and it would be wrong to instate you as a mentor if these are your goals.

Usually the sponsor or the podling by himself ask for any additional mentors (either in the proposal or later on); this then has to be ratified by the Incubator PMC (partly because the mentor will also become an Incubator PMC member).

Cool, so this is the standard +1 from 3 people to approve a mentor for a project?

Yes, where '3 people' == 3 IncPMC members.

An ASF member can also volunteer to be a mentor but IMHO we have to be very careful here that this doesn't get used as a short- track to committership -

Definitely agree here. It should probably be separated really. You can be a mentor but I don't see why that gives anyone commit privs. Mentor or not, if you haven't contributed any code then you're not a committer. This happened prior to the Incubator and I know I was listed as a committer on the projects I helped come to Apache but I wasn't a committer. I think that's easy to fix by applying rules for committership across board regardless of additional roles a person involved in the project might play.

Exactly. IMHO it can be reduced to the following:

new committers = decision of the Podlings PMC

new mentors = decision of the Incubator PMC, nice to have a buy-in from

              providing sufficient oversight

the Podling but in the end the Incubator is responsible for

IMO a mentor shouldn't ever touch any podling code while

wearing his mentor hat; this is totally different from being a committer. In the end it's solely the podlings decision so the Incubator PMC shouldn't impose any new committers...

I don't see why you can't be a committer and mentor. You say it's bad for a mentor to code but you don't say why.

Simple example:

I think you misunderstood me here; that's not what I'm trying to say - I'm very fine with one person wearing both hats; it's just the question when to use these hats. When the IPMC decides that you are a mentor for X then you don't have the permission to go in and change X's build-system to Maven2. You are there to make sure that the project is learning the Apache Way (yah, yah). You are not there to raise any technical discussions which have nothing to do with Incubation. Period!

I definitely think the poddling should have the last say about who comes aboard whether it be committers or mentors.

Cheers,
Erik

I agree for committers and that's what I meant and tried to say with: "In the end it's solely the podlings decision so the Incubator PMC shouldn't impose any new committers...". For mentors this is a bit different, see above.

smime.p7s
Description: S/MIME cryptographic signature