I think we should consolidate to one discussion forum, rather than two.* I don’t think we do a good job of cultivating a community feel, even though everyone in both forums is very helpful and welcoming. It’s hard to remember who’s who across two sites and two avatar systems, and it’s hard to remember to check two places.
As such, here is a one-question survey on which platform you like better.
There are extra questions if you want to provide more info.
In the last month:
7 threads in in github discussions
8 threads in discourse
IMO, the winning platform should have:
Above 50% neutral/positive rating from Beets contributors
Above 40% neutral/positive rating from non-contributors (they presumably will use the forum less often)
If the results between Discourse and Github Discussions are statistically similar, I think we should either flip a coin or let core contributors make an executive decision. (I hope the results are strong enough we don’t have to resort to stats, since I’m rusty on that.)
* Expanded:
Beets has what I would consider to be a chronically small contributor/user base, relative to its usefulness and other open source projects. I think laying this foundation for a strong community could allow it to grow and get more contributors and retain them longer.
Anonymized results here (no usernames, no timestamps):
Note: The final response is mine.
No competitor is as good as Discourse/Github Discussions
There’s no use rocking the boat; because both plausible and implausible replacements for the two main platforms don’t have any kind of support in the community. Reddit is the closest, with about half the (strongly like + like) totals that Github or Discourse does.
Discourse is great! Please consolidate the discussion forum there.
An IRC channel could be bridged to a matrix channel - best of both worlds!
beets is an open source project (yay!), it should use and support open source software for communication and infrastructure.
9 of 13 respondents were contributors
Self-identified, again. A large number (6ish) supplied usernames which we can verify if needed.
Methodology
I use “love” and “hate” in spots as a shorthand for “strongly like” and “strongly dislike”. There is survey theory research that says these terms are not interchangeable.
Survey Design Issues
I realize now that I should have instructed people to leave the answer blank if they had never heard of something. But “neutral” is a similar response.
Sampling Bias: The links were only posted on Discourse/Github Discussions
If someone truly hated those sites, they never went to them and they never answered the survey. With many community projects heading to Discord and Reddit, I wonder if there’s a certain kind of (younger?) contributor we aren’t getting now. Perhaps we could link to the survey in the release notes for an upcoming release, or on the wiki.
Recommendations
Get buy-in from project maintainers that the discussions should be unified. I’ve made my case above. edit: Even if you think we should have multiple discussion locations to bring in different people, I think Github and Discourse are bringing in the same people. I’d sooner support a Github+Reddit combo than Github+Discourse, for example.
Do a larger survey in a more public way. Possibly shorten the survey to just the two main ones and 1-3 of the most plausible others (Matrix, Discord, Reddit, imo). Gitlab is probably a fine platform but I don’t think we’re losing any contributors who would contribute on Gitlab, but not Discourse/Github.
If there’s still a toss-up, convene the most active contributors and have them decide. It doesn’t seem like there’s a lot of hate for either platform, so there shouldn’t be much problem consolidating. I personally won’t care much either way.
