02:42 <JacobWeisz[m]> Hey Ian, is this what we were seeing today: https://blog.domenic.me/hacktoberfest/

04:18 <isd> Probably.

04:22 <isd> Does this mean we're in for a whole month of this?

04:26 <JacobWeisz[m]> I flagged those two PRs as invalid, which is a label Hacktoberfest recognizes as a maintainer indicating a spam PR.

04:30 <isd> I reported them to GitHub; hadn't realized this thing existed at the time.

04:31 <isd> Fortunately we're not quite as popular as the html spec; hopefully the workload won't get too bad

04:31 <isd> But this is dumb.

04:32 <JacobWeisz[m]> They are apparently working on fixing the nightmare they built according to some folks on Twitter.

04:34 <JacobWeisz[m]> I think the problem with this and the Keybase crypto giveaway and stuff all comes down to overconfidence in automation. Security and moderation actually don't scale well.

04:34 <JacobWeisz[m]> DO could fix this easily by having people submit requests to be considered for a shirt and having an employee review the PRs they submit, but of course, that's a lot of man-hours.

04:45 <isd> I think requiring the PRs to actually get merged would do a lot to discourage spam. But also, what's the point of all this? I think there are some baseline assumptions that need revisiting.

04:47 <simpson> Pretty sure it started and stopped with "t-shirt cannons are fun"

10:17 <JacobWeisz[m]> Their response to not requiring merges is that some big projects don't use GitHub to merge, so for those, closed PRs may represent actual accepted contributions.

13:47 <TimMc> Wow, gross. I pinged someone I know at DO and another at GH.

13:49 <TimMc> JacobWeisz[m]: ...which would be so easy to fix with requiring that the label "accepted" be added in those cases, but they obviously want to put all the burden on maintainers rather than "contributors".

14:07 <isd> You could also look at the commits that land, rather than the pr itself.

14:11 <isd> Also, big projects are going to be the ones who suffer most due to the spam, which brings us back to "this whole thing is stupid and they should just not do it."

14:18 <TimMc> From the Hacktoberfest discord, it sounds like if you mark something as invalid, it will get picked up within 2 hours by DO.

14:19 <TimMc> So you could mark all new PRs with "invalid" along with a friendly "ignore the label" note, then unmark later.

14:22 <TimMc> I'm also encouraging Hacktoberfest participants to submit lots of PRs to https://github.com/digitalocean

14:29 <isd> Probably the general strategy should be "help make this make them look bad" if the goal is to make them stop.

14:33 <TimMc> Or go through their partners. I know someone at CircleCI as well...

15:07 <TimMc> Amusing update: https://hacktoberfest.digitalocean.com/hacktoberfest-update

15:07 <TimMc> "Since we likely can’t change the behavior of these participants entirely, we are creating a shortcut for these participants to get their shirt and move on, without having to spam repositories on GitHub."

15:09 <simpson> Yep. My full thoughts are at https://lobste.rs/s/epbcho/digitalocean_s_hacktoberfest_is_hurting#c_o5g8mz but basically I'm imagining how GSoC and EVoC should learn from this.

15:20 <isd> fwiw the xorg page on EVoC doesn't cite spam/low quality submissions as a problem with GSoC. Though I have heard of that being a problem before.

15:22 <simpson> EVoC avoids it by hand-selecting students and having a long and involved vetting process, as opposed to GSoC publishing a Google Docs form for anybody to fill out. But it's not perfect.

15:41 <abliss1> wow, I just skimmed through https://spritelyproject.org/ ... certainly looks ambitious!

15:43 <simpson> Yeah, dustyweb is very productive.

15:46 <dustyweb> hi