Continuing the discussion from Git.snowdrift.coop: 259 likely-spam user accounts:
Initial thoughts (but primarily opening this topic for anyone else to weigh in):
- git.snowdrift.coop is GitLab CE
- we can’t get inadvertantly locked-in to using some proprietary feature (but we also don’t have access to those features if they would be worthwhile enough)
- the strongest FLO purists extra-approve (but note that even GNU/FSF/RMS find GitLab.com acceptable as a host, just not perfect)
- There’s some benefits along with the costs of managing our own overall instance settings and userbase etc.
- GitLab.com reduces the barrier to participation for those with accounts there already
- I don’t know the various other trade-offs in terms of what is more or less functional via GitLab.com such as the way groups and repos work etc.
- There’s definitely somethings nice about having all our infrastructure via Snowdrift.coop TLD…
- I think GitLab pages would be an option at GitLab.com and aren’t for our current set up, but we have many other ways to easily just put up some page (wiki, here at Discourse, our own hosting, ghost/blog, etc.)
wolftune's view: topic is open, but doesn't deserve much attention now
In the end, I’m not going to insist or veto etc. I have my thoughts, am undecided, would defer to others if a thorough evaluation were done, but don’t see the need (and certainly see risk and cost) in putting much time and energy toward even discussing this.
The status quo is quite functional and not a blocker to any of our real priorities, so worrying about it now seems mostly to be a distraction (but it’s okay with me to have this topic open, if only to document things and collect perspectives on this, give newcomers a chance to see this discussion instead of bringing it up again, etc.). It’s certainly possible that I could feel it justified to make the move.