On 11/14/13 9:07 PM, "Marvin Humphrey" <mar...@rectangular.com> wrote:
>On Wed, Nov 13, 2013 at 10:47 AM, Alex Harui <aha...@adobe.com> wrote: >> I still think that having a "Release Auditor" role provides backup for >> getting incubator releases out without having folks have to be on the >>IPMC >> to approve the legal aspects of a release. Just like any ASF Member can >> backup busy PMC Chairs for some actions, any TLP PMC member should be >>able >> to backup a busy IPMC member for release auditing. > >Speaking as someone who would presumably be suitable for this "Release >Auditor" role, I'm opposed to the idea -- and not just because I don't >want to >get stuck doing all the dirty work. > >People who sign up to Mentor a podling should expect to vote on releases >-- >especially the first. The Incubator PMC tasks Mentors with overseeing >the IP >clearance processes. A Mentor who votes +1 on the first incubating >release is >implicitly affirming that IP clearance was done properly -- because that >was >their assignment, and if something had gone awry they would surely not >vote to >release. Well, sure, clearly a highly-engaged mentor can better manage IP clearance. But is release voting really an approval of IP clearance? I thought it was more about IP "maintenance": making sure that everything in the package has a header. Usually there is a significant amount of time between the incubating IP hitting the repo and it being offered for release and I thought the clearance had to happen when it hit the repo, not at release voting time. > >A +1 vote from a "Release Auditor" who did not participate in IP >clearance is >much less meaningful: all it tells you is that whatever superficial >inspection >they performed on the finished product did not reveal any defects. If >some >committer mistakenly attaches an ALv2 header to a file that shouldn't have >one, a "Release Auditor" won't find that. To catch such problems, you >need >someone monitoring the the dev and commits lists: possibly a Mentor, >ideally a >project contributor. I thought the main point of this thread was to find a way to unblock podlings looking to release but their mentors dis-engaged, even temporarily. Are you saying that the IPMC members who step in to help (like the ones who recently stepped in for VXQuery) must do the forensics of IP clearance by scanning the commit emails? Seems like folks doing "release auditing" can do that as well if that's really required. We might even make a tool that searches through repo history for add/remove of copyrights. > >The most meaningful +1 votes are those cast by enlightened core >contributors, >because they speak from deep knowledge of the code base and its history. >IP >stewardship is a continuous process, and the Incubator's goal should be to >graduate communities with the motivation and expertise to attend to it >over >the long term -- not to certify code. Agreed. The only purpose of having a Release Auditor role is to expand the pool of folks who can vote on a release without requiring them to become full-fledged IPMC members. Now if you're saying that having backup voters is not going to meet some requirement of IP safety, it seems like it can just be made a requirement of a backup vote to do whatever that work is. If you're saying that will never work because the only folks who can validate a release are folks who are engaged in the podling, then even having other IPMC folks backup them isn't going to work either, and solutions need to be found to somehow get those mentors to find the time to meet their obligations. -Alex --------------------------------------------------------------------- To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org For additional commands, e-mail: general-h...@incubator.apache.org
