I have a dilemma : I'd like us to take part in google code-in ( https://codein.withgoogle.com/ [high-school level program organized as a contest where we setup a large number of "tasks" ~2h each of various difficulty that students tackle. Could be code, doc, outreach, etc.]) because basically I've heard that it induces a boost in participation and attracts young long-term contributors to participating projects; however it is a *significant* investment : basically around 100hours/week on mentors (shared, not 100h each) during 8 or 9 weeks. The comments from #gsoc when I asked for advice were those : 23:02 < Mc> I'm unsure we (Inkscape) have enough man/womanpower to participate efficiently in google code-in... Do you have a rule of thumb on the number of mentors and of diversity of tasks that would be "nice to have" to try to participate efficiently ? 23:17 < terri> Mc: A few of us were saying elsewhere that we'd want about 20 mentors and 100 tasks prepared in advance, but it depends on how much time your mentors have to dedicate. 23:17 < Mc> ok, 20 mentors is impossible, that clarifies it :p 23:18 < Mc> (100 tasks is clearly doable though) 23:18 < terri> I know people who do it with 3, but at least one of them has no other job for those 3 weeks. :) 23:19 < Mc> Dec2-Jan21 ? 23:20 < terri> *shrug* or whatever it was the year in question. no clue what they're doing this year. :) 23:20 < Mc> I mean, this year it looks like much more than three weeks 23:21 < terri> yeah, that's why I'd want 20 people if I were running it for my org. it's easy to burn out, and lots of folk travel/have family commitments/etc. in that time frame. 23:21 < Mc> (8 weeks ?) 23:21 < fundamental> It's impressive that there are enough open source projects which can deadicate that level of resources 23:21 < terri> maybe I'm just mis-remembering how many weeks it was, though. 23:25 < terri> Mc: That said, that's *my* rule of thumb for a pretty wide umbrella, where we'd need different people for differnet parts of the project. If you've got a smaller subset but they're able to handle more tasks and are more available in that time... well, as I said, it *can* be done with 3 if you're willing to commit to a very high level of activity! 23:26 < terri> It might be better to try to guesstimate how many hours people can put in rather than how many people. 23:27 < terri> 20 people x 5 hours/week versus 10 putting in 10 hours each, or some mix of both. 23:28 < Mc> Ah yeah I can probably commit 3 hours a day most days 23:34 < Mc> 100h/wk would be your rough time estimate (after the first week(s), I guess) ? 23:35 < terri> yeah, something like that. you want "someone's going to be able to evaluate any student's task within a couple of hours" coverage 23:36 < terri> Since it's a contest, students get super nervous if they have to wait too long for task evals, and they can't accept a new task until their existing one is okayed. 23:38 -!- mode/#gsoc [+o meflin] by ChanServ Day changed to 12 oct. 2019 08:56 < PulkoMandy> hi Mc, I think the 20 people and about 100 tasks estimate is correct. However, remember that a part of this can be non-technical (both the tasks and the people). In our case at Haiku we have maybe 3 or 4 core devs mentoring in GCI, and most of our mentors are users or former GCI contestants who are not otherwise involved with the project, they do the 1st level support to the student and know when a particular task needs a more technical 08:56 < PulkoMandy> person to review it. So basically you may need to be a bit creative with how you find mentors for this I think if we include vectors and usual community members we can get to the right number of involved people ( Me, Jabier, Tav, You, ede123, nathan, RdH, NPJ2000, Thomas Holder, CRogers, Maren, Tim, Michele, Mihaela, Ryan, Tedg, Lazur, Brynn, maybe Scislac and bryce ) to not burnout anyone (but me :P) and get a decent number of tasks across the areas of the program (code; documentation/training; outreach;QA;Design) but the bad thing is the timing : 8 weeks in Dec+Jan is basically when we wanted to focus on pre+post-release stuff so it's very badly timed for us (and it's yearly so if we want to try, it would be in a year); what do you think ?