Too many lords, not enough stewards
Too many lords, not enough stewards
Posted Feb 2, 2018 2:51 UTC (Fri) by rodgerd (guest, #58896)In reply to: Too many lords, not enough stewards by neilbrown
Parent article: Too many lords, not enough stewards
This is a great success story, but it seems predicated on someone like yourself who is fundamentally interested in making things better; in this case with the assistance of a third party. You were prepared to recognise things weren't going well and work with the mediator (which is great!); but this doesn't help with the other part of the problem Daniel describes, which is the smaller group of folks who enjoy being able to behave badly - imagine James ringing up someone and being told that their opinion isn't relevant, for example.
(It also seems like the maintainer has a very ill-defined status: are an architect, setting the direction of their subsystem? Are the a leader, focused on getting people to work together? Are the a release manager, deciding when code is ready to ship? All of the above? A little of each?)
(Log in to post comments)
Too many lords, not enough stewards
Posted Feb 2, 2018 3:24 UTC (Fri) by neilbrown (subscriber, #359) [Link]
I doubt Linus, or any upstream, is interested in working with a (sub)maintainer who isn't "fundamentally interested in making things better". Don't suffer in silence with such people. Take specific issues to their upstream ... and be prepared to become the maintainer yourself!!
> which is the smaller group of folks who enjoy being able to behave badly - imagine
I have never seen evidence of those people. Maybe I just lead a sheltered life.
I have certainly had interactions with maintainers who seem to ignore me or seem to be inconsistent, or seem to make unreasonable requests for irrelevant changes. With patience and perseverance progress can still be made. I sometimes get impatient, but that is my weakness.
I won't claim the the group you mention doesn't exist, but I will claim that it isn't helpful to assume any given person is in the group. There are lots of other explanations for apparent bad behavior than the assumption that it is deliberate. If other avenues fail and the evidence for the behaviour being deliberate mounts, then you can use it in your attempts to route around the maintainer.
> It also seems like the maintainer has a very ill-defined status
This is certainly true, and it is a question that various people have tried to address. The lack of clarity is probably one cause of the misunderstandings that arise.
I see the job of the maintainer as simply to do whatever needs to be done - to care about the code, to want bugs to be fixed, to want features to be implemented, to want maintainability to be improved. Those wants will lead different people in different situations to head in different directions.
Too many lords, not enough stewards
Posted Feb 2, 2018 4:34 UTC (Fri) by error27 (subscriber, #8346) [Link]
But last year a top maintainer tore me a second arsehole because I of my commit messages. He complained four times that I hadn't read SubmittingPatches and couldn't follow directions, that I was generally and idiot and that I said, "NULL dereference" instead of "NULL pointer dereference". His final act was to edit my commit to leave a snide message that he was the only genius at writing commit messages and had to do everything for the rest of us. A couple months later he gave a talk about commit messages and LWN wrote an article about how important it is to write great literature in your commit messages. Everyone was like thank you genius man for teaching us with your wisdom.
I am still traumatized by it. Whenever I see his name in get_maintainer.pl output I just start cursing and my day is ruined.
One time I was going to send a bug fix to this dude's subsystem. It was a simple thing where he had returned positive ENOMEM instead of negative by mistake. I wrote the patch and then I saw that there is no mailing list for the subsystem, it's just LKML. I was like FFFFFFFFFF.... No. That's like being alone in a room with your abuser. I deleted the patch.
Too many lords, not enough stewards
Posted Feb 2, 2018 17:33 UTC (Fri) by tbird20d (subscriber, #1901) [Link]
Information will be kept confidential. Just FYI, TAB members reviewed Daniel's talk, and began a discussion on our e-mail list and in our last meeting about the issues raised, and what actions could be taken to improve things.
Speaking personally, and in generalities, there were some things I agreed with in Daniel's talk, and other parts where I thought his analysis was faulty. One thing I found frustrating (but completely justifiable) was the lack of specific examples by which to judge for oneself the merit of his arguments. Unfortunately, when talking in generalities it is all too easy for different parties to think of their own examples, and thus not be able to agree on the nature or severity of a problem. This is hard enough even for people with different backgrounds looking at the same examples! (And, I do recognize the irony that I'm speaking in generalities, and so am likely to not be understood as well as I might.)
This is why the collection of specific instances of the behaviors at issue is important.