Just a comment on discussions: They already have answered/unanswered filters and they have most of the same properties that "stack overflow" questions have,
You do not need to "track" discussions. It's great to answer and react quickly and if you have more discussions all the community might get more involved and start answering. It happened for us after about a month/two of using discussions. The important thing is that "discussion" is a discussion - if it gets no answer, that's perfectly fine - means that the discussion did not pick anyone's interest. Author can still follow-up, ping other people etc. but there is no "expectation" that discussion will reach a conclusion - it can remain unanswered forever and simply disappear. Also this is no coincidence that discussions have no "total count". They are meant to grow "forever" unlike issues, the discussions are meant to just "be there" - sometimes with, sometimes without answers. You can see the discussion in the last day/week/month or search them via keywords but this is about it - there is nothing like "x discussions opened". This is what makes the a fantastic counterpart to issues because you can convert issues to discussions (and back) as maintainer/committer, when you see that you miss information, or that it's unclear whether this is an issue but you have no idea what to do next. They might simply "go away' if the author and others are not interested - or if more information is available or if someone else has similar observation and chimes in it can be revived at any time. But the great thing about discussion it does not leave you with the impression that you have such a big number of "open issues" that are unhandled. Sometimes leaving the discussion open is the right "final state" for it. I did not realize that when we first started to use discussion but "convert issue to discussion" is the single best feature of GitHub issues for me. It does not really "close" the issue (which might be seen as rude and you have to have strong arguments to close an issue), but it gives a clear information to the author and whoever is looking at the discussion that it needs extra effort, clarification, digging (usually from the author but maybe from other interested parties) to qualify it as real issue. We went down from ~880 to 814 opened issues over the last month or so (and we continue our downard route in Apache Airflow) once we made it a bit more difficult to enter the issue (via detailed issue template) and started to promote discussions in the templates and started to actively convert issues into discussion when they qualify as such, J. On Thu, Sep 30, 2021 at 4:04 AM Weston Pace <weston.p...@gmail.com> wrote: > +1 for issues because I believe it would lower the barrier for entry. > > I'm +0 on discussions, they can work but would require more active > curation / labeling as they cannot be closed so an "answered / > unanswered" label would probably be needed. > > > I think I already get e-mails from issues but > > have them filtered out with the rest of other github messages, I'm not > sure > > if it is easy to split them out. > > Issues will absolutely be lost in the flood of notifications you would > get from watching the arrow repo. However, you can do a custom watch > that targets only issues. This may be an alternative for those that > prefer an issue-like workflow. For me personally, I've monitored > issues in the Zulip feed for Github. That being said I went ahead and > turned on an issues-only watch to try that out. > > > I took a few minutes to browse the archives [1]. It seems to me that > > the user@ list is working extremely well. People get answers quickly, > > problems are converted into JIRA cases, and the discussion often > > references existing information sources. > > I would add we have a pretty decent traffic rate for github issues > today. We get a fair number of issues opened even though our issue > template says "Please ask questions at u...@arrow.apache.org". > > On Wed, Sep 29, 2021 at 9:37 AM Julian Hyde <jh...@apache.org> wrote: > > > > I'm not for or against this proposal. > > > > I took a few minutes to browse the archives [1]. It seems to me that > > the user@ list is working extremely well. People get answers quickly, > > problems are converted into JIRA cases, and the discussion often > > references existing information sources. > > > > I want to thank all of the community members who answer questions. No > > doubt it takes considerable time and effort. > > > > Julian > > > > [1] https://lists.apache.org/list.html?u...@arrow.apache.org > > > > On Wed, Sep 29, 2021 at 2:14 PM Phillip Cloud <cpcl...@gmail.com> wrote: > > > > > > On Wed, Sep 29, 2021 at 3:08 PM Antoine Pitrou <anto...@python.org> > wrote: > > > > > > > > > > > Le 29/09/2021 à 20:51, Micah Kornfield a écrit : > > > > >> > > > > >> Cons: > > > > >> - Github is a not a mailing-list and does not integrate well in a > normal > > > > >> e-mail workflow. > > > > > > > > > > > > > > > Would a mailing list mirror of the issues work for you (I guess it > would > > > > > require an extra click). I think I already get e-mails from > issues but > > > > > have them filtered out with the rest of other github messages, I'm > not > > > > sure > > > > > if it is easy to split them out. > > > > > > > > If there's an e-mail notification to user@ (or another place) > whenever a > > > > new issue is created, containing the full issue text, I guess that > would > > > > work. > > > > > > > > > > I was under the impression that you can reply to a GitHub issue > directly > > > from email, as long as you subscribe to issues for a repo. Is that not > the > > > case? >
