On Mon, Oct 31, 2022 at 2:57 PM Shaun McCance shaunm@redhat.com wrote:
Hi all,
You might be aware that there is a CentOS category on the Fedora Discourse instance:
https://discussion.fedoraproject.org/c/centos/71
There have been some discussions about having a dedicated Discourse instance for CentOS. Discourse has a lot of advantages, such as better moderation and integration with our accounts system. But I understand that many people are comfortable with their existing workflows.
There's no point in running Discourse if it doesn't get enough buy-in. So I'm asking for input on using Discourse for various things:
Project announcements, like events, meetings, and infra changes
Activity reports, such as for SIGs and events
User support, replacing forums.centos.org
Development of Stream itself, basically centos-devel
Development of stuff inside SIGs
Replacement for the comments section of the blog
Alternatively, just replacing the blog entirely
Something else I'm not thinking of
From my point of view, I'd look at a CentOS Discourse as a replacement
for the older CentOS Forums. I would rather not replace the developer discussions with Discourse, but user support and engagement places, sure.
Activity reports and such should be going out to the blog because that's how the media is going to pick it up. Notably, the CentOS Hyperscale SIG is continuously in the news because we did the extremely simple thing of always having our reports on the blog. SIGs that don't do that don't get talked about. They don't get mindshare, and they don't get growth and further interest.
I prefer the mailing list for developer discussions because it allows me to tag people into discussions easily enough. However, CentOS Stream development is currently not in a very good place because almost nobody from RHEL engineering is here. Same goes for the IRC channels and any other medium. CentOS Stream development remains horrifically opaque, and that is a bug. Unless things change at some point, most mailing lists could be closed with not that much impact, since there's no communication anyway.
(As an aside, the amount of backchannel effort I have to do to even get stuff to be *looked at* is pretty awful. If we want this to be a successful project, the mindset of how people are supposed to work with RHEL developers needs to change.)