https://docs.all-hands.dev/ logo
Join Slack
Powered by
Hey all, I wanted to get some feedback on how we ...
# feedback
e
Hey all, I wanted to get some feedback on how we might be able to improve our community processes for Feature ideas & brainstorming requirements. Awhile ago, I was reading about how other OSS Projects manage contributions, and came across this note from Mitchell Hashimoto about how they use GitHub Discussions vs. Issues. I was wondering if other folks have worked in other Open Source communities where GitHub Discussions was used as a place to sort through Feature requirements/acceptance criteria before an issue was created, and if it was helpful (provided contributors with clearer direction) or had other side-effects. Thanks for any feedback.
b
From my experience either teams use that or purchase a tool that does the same. I think there's other OSS versions out there but Discussions is probably the best considering.
c
Thank you for opening this. I'm not sure about introducing… a 3rd place for issues (3rd? GitHub Issues, Linear, GitHub Discussions?) I think maybe Automations and labels will help with triage 🤔 I believe we need to use the agents more… Human attention is the bottleneck, I don’t know, no matter if stuff is in issues or in discussions. As my cat would say, “humans, they’re cats who can’t jump on all things pfft” - or something to that effect, sorry for being silly 😹
e
Sometimes there are multiple issues created for same feature, or the issues have overlapping and/or conflicting requirements. But those folks don't see those issues so end up creating PRs that then get held up. This isn't always the case, but I think it might happen sometimes. I'll see if I can get examples.
c
nod I hear you, what do you think of something like this: finding duplicate issues?
🙌 1
p
I love the idea of leveraging agents for finding duplicates (and experimenting with how they can help with triage more generally)!
I'm not sure about introducing… a 3rd place for issues (3rd? GitHub Issues, Linear, GitHub Discussions?)"
I think this would actually mean we just have one spot for introducing potential items to work on, and that would be GH Discussions? Those would then be converted to issues once it is ready to be worked on (per the note Joe sent above). Linear would remain simply a mirror of GH issues, used by the internal team to track work. Assuming that is the case, and we're just moving the spot where potential work items can be created instead of adding another one, I'm curious how people feel about that? Pros and cons?
c
Just to note, it’s working 😅 • duplicate issue auto-closed • issue with thumbs down (should remain open) • issue with duplicate information, before auto-close
🙌 1
I appreciate your thoughtful response, Ash, and I see how you mean, but I don't fully understand one thing. I hear in Joe's words and yours, that you are trying to address a problem on OpenHands/OpenHands repo. I know, AI PRs are increasing and people are getting tired. But that seems like solving the wrong problem, a problem that doesn't need to exist in the first place. Since January, we have planned to move Web <-> OH <-> SDK codebases. That solves the problem for the Web Server repo. Because AI PRs are attracted to the popular repos on GitHub, they will rarely come on other repos. Many are semi or fully automated, agents read stars on repos, they clone 20+ repos in new accounts, and shoot PRs. Often, without a human using OpenHands as a human user even once. On the other hand, we are moving the SDK in, which of course is what OpenHands has always been: Open Source software, under MIT license, designed to be the best AI Software Engineer we are building. De-duplicating issues, understanding PRDs, implementing, reviewing, testing, is exactly what we teach the OpenHands agents to do. For us, those working on that, AI PRs are a gift, not a problem. Data! 😅 I do feel the urgency here. I really do. I believe that from many points of view, not one: we need to finalize the repos layout issue. Finalization: • We have made the OpenHands-Web repo, and it's simple; it's exactly the contents of the current OH/OH • I have tested the merge of histories of the SDK into OH/OH, and it works like a charm • I have implemented an issue forwarder from OpenHands-Web to the historical issues on OH/OH • and live tested it on real issues: it works. Left to do: • building a release from OpenHands-Web • deploy-testing on Saas • move • I'll run the forwarders, PR notifications, other minutiae. It looks like an absolute necessity to me, and it's a win/win. We solve the right problem. Then, there is no need to solve other problems that are not, actually, problems. WDYT?
e
Thanks Engel! Re: repo changes, I think there's interest from the other maintainers in moving ahead with moving repos. I am not as closely involved in it, but I think it will help some problems. Re: using AI/OpenHands to help with identifying dupes, that's a great idea and something we might want to adopt! My original post here was mostly looking for feedback from others who have used Discussions as a way to align on requirements before Issues were created. That was my main intent here, just seeking feedback 🙏
2 Views
Open in Slack
PreviousNext
Powered by