* Resetting the list of committers to people that have contributed in the previous terms in the latest 6 months
I bought this before and got shot down by the mentors, as according to Apache it is impossible to bar merit based on time of contributions. On the other hand, I like some of the points being raised in the proposal. Note that, the "bar of engineering" is something that can be discussed in the dev-list. Software engineering is about trade-offs. We are already writing test cases, document APIs to prevent as Pedro suggested in the list. On the other hand, inevitably sometimes bug slip through and the rate, especially when there are major changes. Having flaky test cases and bug appearing should not be the absolute reason to accuse bad bar of engineering. Because there is a trade-off to make, it is better to delegate some of the trade-offs to the committers who do the code reviews and encourage more contributors to contribute code reviews. It is great to raise concerns about specific changes and say how things could have been improved to give more context. But in many cases, we should also know that our opinions are biased especially if there is a specific area we tend to. Some of us are more like "protectors" who would tend to shoot down contributions that may not have "evidence" that everything is covered. Some others are more like a "pioneer type" who would try drastically new designs while ensuring test coverage, sometimes brings in bugs. Anyhow, when the decisions are made, it is made by a group of people, everything is helpful but nothing is perfect. CI, test-cases and even one vetoing code reviews cannot judge the status of the contribution. I usually find it is most helpful to still delegate to the committers who are familiar with the code to make the final decision. Tianqi On Sat, Jun 16, 2018 at 8:57 AM, Pedro Larroy <pedro.larroy.li...@gmail.com> wrote: > Hi Sebastian. > > Thank you for your comment. That's why I said "I would propose", because I > don't know if it's possible as my experience with Apache is limited to the > MXNet project. > > How do you interpret this part?: "Since the appointed Project Management > Committees have the power to create their own self-governing rules, there > is no single vision on how PMCs should run a project and the communities > they host." > > Pedro. > > > > On Sat, Jun 16, 2018 at 8:37 AM Sebastian <s...@apache.org> wrote: > > > > I would propose the two following action points: > > > > > > * Resetting the list of committers to people that have contributed in > > the > > > previous terms in the latest 6 months > > > * Suspend veto rights temporarily and use simple majority for > decisions > > > that need a vote so we realign the project with the communities best > > > interests. > > > > None of this is possible within the structure of the ASF, please get > > familiar with how the ASF works before making such proposals. > > > > https://www.apache.org/foundation/how-it-works.html > > >
