[sdesol]

> False Assumption #1: Geographic Location Doesn't Matter

I can't find it, but maybe somebody from Microsoft can, but they (Microsoft) actually found building distance made a difference in productivity. This was in a research paper that I read, but for the life of me, I can't find it. I should also add a disclaimer that, this research paper was done before remote work became more of a norm and we had less technical options.

> Night owl behavior is actually the exception, not the norm.

I'm currently not tracking hours that people work with my developer analytics solution, but I think if would be flawed to just take into consideration when a pull request is authored to gauge behaviour.

If you look at the pull request information for vscode (my goto project for good data points) at:

https://public-001.gitsense.com/insights/github/repos?q=wind...

you can see that for some pull requests, there can still be quite a bit of work from when a Pull request is authored to when it get's merged.

[musingsole]

I submit pull requests based around a few hours throughout the day my team seems most likely to be between tasks (and so likely to check out a PR). Such as around 11 or 2 for pre/post lunch ramp down/up. Point being, the time my PRs are submitted has no bearing on when I did the work for that PR.

Using repo metadata to arrive at productivity metrics always strikes me as willfully bullheaded. If you timestamp my keystrokes, you can't know when I designed the algorithm I'm coding with those keystrokes. Spoiler alert: it was probably while I was falling asleep the night before.

[sdesol]

I wouldn't go as far as calling it "willfully bullheaded" as I do believe knowing when somebody creates a pull request can provide some data points worth mulling over. Having studied hundreds of popular open source projects, there does seem to be a pattern as to when people prefer to create a merge request, which is mid week.

I do agree that GitPrime, GitHub Insights and other similar solutions are pushing developer metrics in a dangerous direction, by latching onto low hanging fruit metrics. I written a bit about what I believe is a positive direction and this is focusing on impact, which I talk about at

https://news.ycombinator.com/reply?id=26457072&goto=threads%...

[mprovost]

The classic study of how distance affects communication was done by Thomas Allen in the 70s and results in the Allen Curve [0] which shows that people are four times as likely to communicate regularly with someone sitting six feet away as with someone 60 feet away, and that they almost never communicate with colleagues on separate floors or in separate buildings.

[0] https://en.wikipedia.org/wiki/Allen_curve