having some thoughts on the mailing list workflow....

i find that almost every time i email a patch there's some stupid simple issue with it that has to be addressed with a new revision...

my attention to detail when it comes to patches is really bad, for sure a part of it is lack of experience, but I'd be remiss (haha) to pretend that ADHD doesn't play a part. I really struggle with the whole checklist thing, making sure i do everything i need to before sending.

When contributing via a git forge, versioned patchsets are replaced by a source and target branch, and a log of force-push history. This means that when I spot these stupid mistakes I can just fix them. It drastically lowers the barrier of entry for me and reduces the impact both on me (not having to be super self aware of my stupid mistakes) and maintainers/reviewers (not having to point out stupid mistakes that i already noticed myself immediately after sending).

idk, im sure this is something that improves over time. I just wonder if there's any merit to this being a repelling force for folks who would otherwise really enjoy working on projects like the kernel....

@cas Honestly if anyone suggests using anything email-based in 2023 it is not due to any kind of benefit of the workflow, its soley for making sure that only a small, highly technical ingroup can ever reasonably contribute.

In my entire uni class of like 300 people, not a single one has even considered using anything other than PRs, or has used Email even once for dev stuff

Just the Old Guard™ trying to keep projects in their ingroup IMHO

@pojntfx @cas I have not found any fully distributed issue/patch management system compatible with #Git that has had any adoption, the next closest is #decentralized #email which, if mirrored onto users' machines (as they should be) can also count as #distributed.

#GitBug is such a system but since it has barely any adoption…

#Fossil meanwhile includes all the necessary tooling, but has this annoying and misguided focus on cathedral development style that greatly limits its applicability.

@pojntfx @cas #Email is far from perfect of course, more and more hosts these days require doxing yourself to register, which is utterly unacceptable.

@lispi314 @pojntfx yeah i do appreciate the necessity of having development be distributed with a project like Linux. I also appreciate that changing maintainer workflow is really hard.

I'm not nearly as extreme in my position as Filicia i think xD

@cas @pojntfx For what it's worth, I do agree we need something else.

Email is only slightly more resilient to censorship than full centralization.

I think accepting more centralization is a move in the wrong direction.

Is #GitBug the ultimate solution? I don't think so. It's the main one in development and usable right now though.

The corposcum PR apparatus pushing their centralized #SaaSS so hard hasn't helped, distracting many capable of coming up with decent alternatives from doing so.