[wladh]

While this works for freelancers/students/startupers, it doesn't work very well for people working in bigger companies. If your interests are related to your day job, I think it's quite risky to put your code on GitHub, even if the code is not "official". While I do have a GitHub account with some code in it, that code is not really representative. If having a relevant GitHub account becomes a pre-requisite, you may be overlooking some good people.

[technomancy]

> you may be overlooking some good people.

Every heuristic overlooks good people.

This heuristic has the advantage of having next to no false positives, even though it allows for lots of false negatives. Given that false positives are absolutely disastrous while false negatives are merely unfortunate, it's a trade-off I'm happy to make.

[wladh]

I guess the majority of repos are either bigger and useful open source projects or just small toys. If you're looking at somebody with repos from the first category, most likely his resume will mention "author/ contributor to project X". For the people in the second category, only thing you'd infer is that they can somewhat code. Which you could assume as well if they worked on any delivered project. There are obviously exceptions in both cases.

[heresy]

Yep. Without thinking, I used code for an open source library I had written in my spare time in our proprietary software, and it came up during a due diligence check for acquisition.

Wasn't a big deal in the end, I just made sure it was using a commercial friendly license, but you do have to take care.

If it was a US employer they may have sued me if they didn't think that the OSS code came before the inclusion in the product (which it did).