I'll keep just linking people to my github, and showing them projects I've built.
[1]: Actual real work as in: writing libraries that other people use. Contributing to the community in ways that effect the entire community.
I worry a little about the GitHub approach as well.
On the one hand, it's great that some people have effectively got portfolios now. It's verifiable evidence that they have some clue what they're doing.
On the other hand, I'm concerned about a bias developing against people who don't put loads of work on GitHub for whatever reason, which is not verifiable evidence that they don't know what they're doing. They might be world class experts who could easily demonstrate their skill and expertise in person, but that's no good if employers all start using cost-cutting auto-screening software that never shortlists such candidates for interview because they didn't share their ingenious but proprietary code/ideas with the general public.
In short, any scheme that relies on historical demonstrations, whether it's claims on a CV or code on a GitHub account, is always going to be vulnerable to false negatives.
Any scheme at all is always going to be vulnerable to false negatives. There are companies that have multiple-stage recruitment, with skill tests, psychometric tests, multi-person interviews, background checking - and still end up with a few duds.
But if you can use a bunch of proxies (like StackOverflow, GitHub, and LinkedIn), and know the value of each of those proxies, then you can more efficiently match a person with a role. Sure, there'll be lots of mistakes, but it's hard to see a better alternative.
[1]: Actual real work as in: writing libraries that other people use. Contributing to the community in ways that effect the entire community.
I worry a little about the GitHub approach as well.
On the one hand, it's great that some people have effectively got portfolios now. It's verifiable evidence that they have some clue what they're doing.
On the other hand, I'm concerned about a bias developing against people who don't put loads of work on GitHub for whatever reason, which is not verifiable evidence that they don't know what they're doing. They might be world class experts who could easily demonstrate their skill and expertise in person, but that's no good if employers all start using cost-cutting auto-screening software that never shortlists such candidates for interview because they didn't share their ingenious but proprietary code/ideas with the general public.
In short, any scheme that relies on historical demonstrations, whether it's claims on a CV or code on a GitHub account, is always going to be vulnerable to false negatives.