| 18 Apr 2024 |
 ronef | Also, would you prefer to create a matrix room inside of the Nix channel or do you want to keep it a separate matrix channel completely? (might be a bit more discoverable with the former) | 22:39:29 |
 @reedrichards:matrix.org | Yea I'd for sure prefer to keep the matrix room inside of the nix channel! | 22:49:04 |
| @reedrichards:matrix.org | I'll shoot you an email | 22:49:10 |
| 19 Apr 2024 |
 delroth | question to the board: who from the foundation side is tracking how the Summer of Nix money is being used and whether that's actually a valuable investment for the community? | 16:00:49 |
| delroth | my semi-external view of Summer of Nix at this point:
- almost nothing gets contributed back to nixpkgs - the result of SoN is too low quality to actually contribute back so instead it gets stored to rot in ngipkgs
- the foundation pays people to admin the program who seem to have very little experience with nix development and are more interested in pushing development techniques that haven't proven to be effective ("mob programming" - all I've heard from prior participants is that it's a waste of time, and it's very much pushing away potential participants, to the point that I think some exceptions have been made this year?)
- you could argue it serves to attract new contributors, but uh, do we have any evidence that this works? I know of one prior participant converting to becoming a contributor, but the program really isn't structured well to do this.
- in the meantime, it's the biggest cost center to the project. Yes, it's all paid for by nlnet in the end, but uh, that's not a good way to use public money on nix?
| 16:07:01 |
| delroth | and let's not even talk about the SoN money wastage that gets put in the infrastructure bucket - like the completely idle 300€/month machine that's been in our hetzner account for a year completely unused but earmarked "for SoN" | 16:07:41 |
| delroth | I know ronef we talked about this at FOSDEM this year already so at least this shouldn't be news to you... | 16:08:52 |
| delroth | * and let's not even talk about the SoN money wastage that gets put in the infrastructure bucket - like the completely idle 300€/month machine that was sitting in our hetzner account for a year completely unused but earmarked "for SoN" | 16:11:34 |
 Pol | Funny... I asked a question about that on an issue on Github at https://github.com/NixOS/nixpkgs/issues/237307#issuecomment-2017437870 | 16:17:23 |
| delroth | to be clear re: "mob programming", and to put numbers on things: the 2023 SoN report claims "Knowledge sharing: Mob programming is very effective" while also saying SoN spent 3225.5h at 25.52€/h packaging 5 projects | 16:17:40 |
| delroth | that's one project per month | 16:17:54 |
| ronef | I haven't been personally looking into it but I'm happy to
Agreed that SoN is a critical project that can have lots of good impact! | 16:18:24 |
| Pol | SoN, Mob, <call-it-whatever>, if there's no contribution back, it's almost pointless. | 16:19:22 |
| delroth | Pol: the written documents claim that the aim is to upstream stuff but all the actions from the SoN admins go against that - the existence of ngipkgs is almost on its own an admission that their effort can't actually meet normal nixpkgs quality standards | 16:21:20 |
| delroth | (and like, normal nixpkgs quality standards aren't high, but for example https://github.com/ngi-nix/ngipkgs/blob/main/pkgs/by-name/freeDiameter/package.nix wouldn't pass...) | 16:21:32 |
| delroth | (for very good reasons) | 16:21:36 |
| ronef | What's the biggest problem items today from your perspective?
Seems to me that there's a value problem, velocity? | 16:21:48 |
| Pol | I know, I have been asked for helped in the ngipkgs repo, I was really surprised that stuff were merged as-is. | 16:22:02 |
| ronef | Querying so we can start tackling prioritization of problems | 16:22:32 |
| delroth | ronef: nixpkgs gets ~0 tangible value out of SoN, maybe gets indirect value by introducing some new people to Nix but that's dubious, and all the while it's a giant money pit | 16:22:50 |
 tomberek | The 2021 SoN was run differently and I would say was far more successful in both upstreaming, bringing in beginners, and onboarding long-term contributors. (report: https://summer.nixos.org/assets/report-2021.pdf) | 16:23:15 |
| ronef | (also delroth don't know if it was intentional but really loved that this topic was brought up when I'm awake 😅) | 16:23:31 |
| delroth | it's not intentional, I'm just basically on california time right now after spending all night trying to figure out Hydra :p | 16:23:58 |
| Pol | I'm on walifornia time, so it's fine too. | 16:24:24 |
| delroth | in essence uhhh I feel like SoN reports are making a lot of claims that are trivially disputed by public evidence | 16:25:08 |
| delroth | I quoted the efficiency one above, but that's not the only one | 16:25:16 |
| ronef | I also recall that 2021/2022 had better perceived value? Happy to hear if that was the case in other's eyes as well just as a reference point | 16:25:23 |
 piegames | In reply to @ronef:matrix.org
Querying so we can start tackling prioritization of problems From my outsider perspective, "mob programming" is the #1 issue, as I see it as the cause for many of the described symptoms (low efficiency, poor mentoring, poor contributor retention, lack of Nix expertise in group leaders) | 16:27:43 |
| delroth | to be clear: I wasn't sure if it was just my own experience/feeling or not, so I've talked with several experienced nixpkgs contributors who said they wouldn't even consider mentoring for SoN because of the "mob programming" format | 16:30:14 |
| delroth | * to be clear: I wasn't sure if it was just my own experience/feeling or not, so I've talked with several experienced nixpkgs contributors who also said they wouldn't even consider mentoring for SoN because of the "mob programming" format | 16:30:31 |
