For the past year and a half @kumaraditya303 has been actively contributing to CPython, specializing in asyncio, deepfreeze, and performance topics, also branching out into general C code cleanup and documentation.
He gained Triage privileges this year on April 5, 2022 and in that role he has been incredibly productive.
For asyncio issue triage I now rely on him and often for PR review as well – I have quipped that it’s been annoying to have to do the merges after he’s done the review. (And yes, I do re-review those PRs.
Currently he has 146 merged PRs in CPython. He’s been involved in hundreds of issues and PRs.
I know Kumar’s contributions in the past have occasionally been somewhat controversial – I have mentored him and am confident that his judgment has matured. Nevertheless I intend to continue mentoring him for 1-3 months after he gains core dev privileges.
With his skills and experience, I believe Kumar would be a great addition to the core Python team.
Big +1 from me. First of all, Kumar is a high-class C coder, but he’s also very quickly become proficient in the inner workings of the interpreter/runtime. I’ve learnt a lot from collaborating with him on various PRs, and I always welcome his competent review remarks and insightful comments.
I often don’t write comments about nominees with some feedback.
But it might be the last chance to give him frank feedback.
In conclusion, I gave him a +1. This is because I respect Guido’s judgment and Kumar’s contributions, except for my personal experience of interacting with him. I will not list my experiences with him. This is just my experience, and in fact, it may be nothing after that.
I’ve talked about it many times, his skills were enough to become a core developer, and I thought it was only a matter of time before he became a core developer.
However, I think that being a core developer should give a lot of value to doing code reviews by external contributors rather than writing codes. In fact, I presume that more than half of the core developer’s tasks are code reviews.
Most of my uncomfortable experiences were probably the code review process with him. Whether his intentions were pure or not, basic social skills are necessary, and I think that external contributors who are not core developers are not interested in his personal history and will consider his code review experience as the code review experience of the core team.
So I expect him to show a more mature attitude if he becomes a core developer this time.
Finally, I don’t think age is a prerequisite to having any responsibility, but I do believe that there are some abilities that are necessary to have a certain responsibility. And I expect those things from him.
For more additional comments on his promotion:
But I wish that he becomes a core dev for the following reasons even if the above feedback.
(This is why I gave +1 for his vote)
His deep-frozen project was very impactful for CPython 3.11 release.
His activity with asyncio projects will resolve the lack of resources to asyncio project reviewers.
He is active in the subinterpreter project which I love and he achieve a lot of progress these days.
I believe that Kumar will overcome my worries with his great mentors
+1 without hesitation. Kumar started contributing to cpython while still in high school (I’m not sure if he finished by now?). He needed to learn not only the technical aspects, but also the team collaboration side, which many of us got to pick up in our first paid job after university.
I am impressed by how much he has achieved and contributed in this time, as well as by his commitment and perseverance. It could not have been an easy process (we don’t tend to hold back on feedback…). I’m confident that he will continue to grow and be a valuable member of the core team.
Like others, I have received useful code reviews from Kumar. We all benefitted from many hours of issue triage by him. He is currently our most active asyncio expert.
In my experience, contributors rarely do core reviews because most core reviewers only trust reviews done by other core reviews. When a contributor is promoted, usually they tend to do more reviews than previously for two reasons. First, reviews of core developer are more visible, impactful and respected. IMO the second reason is that once a contributor is promoted, they tend to interact even more and be involved in even more topics. It’s a virtuous circle.
I don’t want to discuss why reviews of contributors are not taken in account here, it’s the wrong place, this topic is only Kumar’s promotion.
+1 from me also. It’s been fantastic to see how much and how quickly Kumar has learned since he started contributing to CPython. He’s matured into a trusted member of the team (and knows far more about asyncio than I’ll probably ever want to).
I do not like to give negative feedback, but I’m strongly -1 to promote Kumar. In my opinion and based on previous interactions with Kumar, I would feel uncomfortable to give him the commit bit now.
I’m not doubting his coding skills. Just based on his C skills he would be an excellent addition to the team. However there are other traits which are as important as technical skills. Core developers have social responsibilities towards the community and their fellow core developers.
It is my impression that Kumar is taking too many risks and is too aggressive when it comes to new features or closing of bugs. The stability of Python is of upmost importance. I think that a good core developer should be humble, careful, know their limitations, and have a strong culture of failures.
Failures happen all the time, no matter how careful and skillful a core dev is. That’s why a core dev must not be afraid to admit, apologize, and learn from their mistakes. Otherwise we end up in situations like former, now banned core developer Stefan Krah. Stefan was unable to recognize and apologize for bugs he accidentally introduced. Instead he leashed out and attacked community members and fellow core developers when they reported regressions in his work.
Kumar’s problematic reactions and feedback around the “asyncio ssl” release blocker are far from Stefan’s attacks and CoC violations. However since then I have not seen any indication that he acknowledged mistakes have happened or apologized for the fact that several core developers had to spent multiple hours debugging the problem. Core developers should neither be too proud or too afraid to say “Sorry” or “Thank you” to a contributor.
Therefore I would prefer to delay Kumar’s promotion.
Christian, I hear you. And thank you for speaking up.
This is one area where I will definitely watch Kumar’s activities and interactions closely and will provide him with feedback and suggestions, regardless of whether he is promoted.
It is granted by receiving at least two-thirds positive votes in a core team vote that is open for one week and is not vetoed by the steering council.
Here 5 core developers voted -1 and 27 votes +1. The 5 negative votes are not enoguh to block the promotion, since the 27 positive votes fit the 2/3 majority rule (84% is greater than 66.7%). But I’m a little bit worried that the number of votes is so high and “ignored” by PEP 13. Before PEP 13, it was common that a single “-1” vote was treated as a veto (block the promotion). Well, the governance evolved for the best and for the worst
Note: I voted +1 but I’m worried by Dong-hee’s and Christian’s messages.
Perhaps this is because (I think) voting used to be public, so there was an expectation for -1’s to identify themselves and provide some reasoning for their negative vote? It’s a lot easier now to just vote “no” for any reason - my own promotion received a single “no” vote, and I still have no idea who cast it… or, more frustratingly, why they cast it, since I only got positive comments on the poll.
…which, by the way, is part of the reason why I’m glad that a single -1 is no longer treated as a veto.
Well, I got -2 votes for my own promotion, I still don’t know who they were and what their concerns :),
But I can ask members of our team who gave me +1 votes, “Did you regret giving me +1 votes by any chance?” and I wish that it would be “no”
so @kumaraditya303, Welcome to the team, and please prove my concerns were wrong. (I am not in -5 vote, but I published concerns to the public.)
