Evan's metric and henry's are not directly comparable.

Someone on vue core team (perhaps evan himself) said internal group co-ordination and management is handled by other team members, most prominent among them being Chris Fritz[1] (before he stepped down) and Sarah Drasner[2].

This reduces managerial burden on Evan and gives him more energy & time to focus on technical challenges

While Henry is the senior most member in babel team, so most of the management work is on him. which eats his time a lot more.

[1] https://twitter.com/chrisvfritz [2] https://twitter.com/sarah_edo/

The necessary "glue" work that often doesn't get tracked but eats a lot of time, energy, and cognitive capacity is very often overlooked. I'm not saying that is or isn't what happened here.

I've worked with a lot of teams where no sort of solid leadership existed. Filling that gap is often difficult to justify because there's a tremendous amount of thought involved and it's a thankless task. Sure you can take the existing structure and direction and just plug away but ultimately, most efforts that do this fail. You need to look ahead and think about future adaptability if you want success. People often act like software development is as straightforward as data entry. You should have X entries per day. If you didn't add X entries then you must be slacking. Development requires not only technical prowess which is difficult to maintain in-and-of-itself but creativity, vision, and strategy. If you ignore the rest your project will fail or at least become an artifact of times past at some point.

We actively discourage team leads from trying to be the "team lead that still codes a bit here and there", they mostly just get in the way and don't focus on the team's real needs.

I.e. we'd rather them take a late lunch or go home early than push a sub par PR nobody wants to criticise.

> it's typical to have a 'contribution' on GitHub every day

What about long running feature branches that get squashed before merging?

To add another data point, on the BabelJS Slack, in the last 12 months, @Henry (hzoo, based on the matching profile pictures) has posted 96 messages, and @nicolo-ribaudo 251.

One can bring the messages up with these filters:

"from:@Henry before:2021-05-12 after:2020-05-11" and "from:@nicolo-ribaudo before:2021-05-12 after:2020-05-11"

contribution in general doesn't have to be "code" related.

You can send a PR with 1 million lines of code, or 3 lines of code.

You can send PRs updating typos in documentation, or PRs affecting the core aspects of a project.

You can send 100 PRs full of tech debt or 1 that makes the thing run 100x faster.

PR counts suck as a way to measure work.

It's the new LoC metric...

it's insane Github reports PR stats. I always knew this would happen. But never so publicly. You just assume that your management is looking at it constantly and basing their decisions on it. Because why wouldn't they. Their entire philosophy is "never let a metric go to waste."

All this does is encourage bad behavior. Not squashing commits, throwing up tons of one line changes. Remember that contest Digital Ocean did for Hacktoberfest? Remember what a total shitshow they caused on Github? People were making egregious commits for a fucking t-shirt. Now imagine real money on the line.

Technically you can do and undo the same thing a trillion times and have a trillion commits. In the end your net impact in the code will be zero, yet you will be the PR champ.