Any career, especially in our field, requires a high level of skill. We try our best to level the playing field for everyone while still getting a lot of signal in the interview process so end up eschewing things like school attended, talks given, OSS contributions in evaluating candidates. Anecdotally we've seen little correlation with these sorts of things and interview ability or ability at the job after being hired.
And even more have the potential to become excellent coders but didn't have the bandwidth to develop it.
It seems peculiar to single out one quality in particular that sends a clear signal that a skill has been honed on the basis that it took time to hone that skill.
All skills take time to develop.
Besides, having a diverse experience base and pipeline (parents, non traditional developers, non traditional paths to SWE, and even single people with a ton of time and fortune to be able to do things like robust side projects) has served us particularly well in having a team with a wide breadth and depth of experience and viewpoints. Specifically weighting side projects in lieu of technical/EQ interview performance would ruin that in favor of the latter group.
Nobody said that we did.
This is about ignoring OSS contributions vs. reading them taking them into account - i.e. deliberately ignoring a signal of quality because it might, for instance, discriminate against people who chose to have kids.
I find it particularly ironic coz part of the reason I wrote open source was to save time - to skip wasteful technical interviews that it ought to be obvious are unnecessary if I have public evidence I can code well.
To your second point, I can't think of a single thing I would let be a proxy for technical skill in an interview process -- certainly not some commits to open source. (Maybe if I had worked side by side with the person in a previous life would be weighted) Even if we did use that as a proxy, it doesn't really save time because people would have to inspect the work and still doesn't really tell me anything about the candidates approach to problem solving or framing of issues that I can get from pair programming or walking through some code for an hour.
I applaud all your open source contributions and appreciate them, but I wouldn't consider them in whether to hire you except perhaps at the margins. Others may disagree and that's their conscious choice or they're indexing on something different.
I did suspect that underlying your opinion was a desire to discriminate in favor of parents/against non parents.
This would fit in with the double standard I highlighted in my first comment.
This was also because you mentioned "parents" in the context of diversity (which is weirdly unique). I threw that phrase out there to see if it triggered you.
>To your second point, I can't think of a single thing I would let be a proxy for technical skill in an interview process -- certainly not some commits to open source
I've not worked with a huge number of developers who have made > 3 significant pull requests to a serious OSS project but every single one has been stellar.
I've worked with a lot of developers who can do the cracking the coding interview dance who sucked and even more who interview well in other ways.
A full time job and small children for example don't leave much space for extended coding sessions on side projects, unless you're lucky enough to need little sleep.
I've worked around this by learning at my day job : read books/papers at home (this is relaxing), and look for opportunities to apply this knowledge at the office. Unfortunately, it stays at the office though, and can't be shown as proof of your skills outside of the company.
Now of course this doesn’t matter past a certain point — people could spend all their time working on something that doesn’t help them grow.
Then once you have those 10,000 hours of actual growth, you are probably close to a Senior developer level. Which after 5 years in an office job working 40 hour weeks, you’re there anyways.
But at the earlier levels of developers, it seems that working on projects outside of work would definitely help you improve faster!
I'm sure a lot of people who hire developers think of themselves as people who could write open source but don't have the time.
This also accounts for why stuff like leetcode persists (the hirers were successfully selected by that process), why top colleges are often preferred and even why the profession is predominantly white/male.
Except that 10000 hours thing is non scientific nonsense.
To improve, you have to practice right way. If you code for 8 hours at work, coding 2 further hours wont make you improve more. Similarly, if you want to improve in music or running, just playing songs or jogging means you will hit plateau pretty fast. After that, you have to train on correct selection of exercises.
At that point reading some theory will have much bigger impact, because yoi are doing something new. And even exercising will likely make you improve more due to what it does to body then further 2 hours of the same.
A) I'm pretty sure almost nobody actually codes 8 hours at work.
B) Those two hours will be of a very different nature to "work coding" and this can lead you to learn things you wouldn't pick up at work.
The non-coding parts of work are necessary on oss projects too. In any case, programmers work being mostly coding + code review is pretty normal.
> Those two hours will be of a very different nature to "work coding" and this can lead you to learn things you wouldn't pick up at work.
That is nowhere near guaranteed and more likely to not be true. In particular, if you focus on maintaining the same software, it will be more of the same after a while.
IME a far greater % is coding. There's also a lot of OSS people involved who don't code who help out with the other stuff.
Moreover, a lot of the stuff that "isn't coding" is a strong quality signal - how the developer interacts with bug reports for instance.
>That is nowhere near guaranteed and more likely to not be true. In particular, if you focus on maintaining the same software.
It'll still be different to what you do at your day job and will necessitate picking up a different set of skills and a different perspective.
It was not my experience. In work, the proportion of coding was bigger then when I maintained open source one. Mostly because company paid people to do other stuff. In OSS, afaik, it is was more of rare to have dedicated tester or to get already analyzed input. Also, the planning, organization, documentation, keeping tickets clean, decision making and so on are all on you. The ratio of project work is the same.
> It'll still be different to what you do at your day job and will necessitate picking up a different set of skills and a different perspective.
Maybe yes, maybe not. And after a while of maintaining the same software, you already maxed out what you learned. Moreover, you can learn new stuff without having side project. It has advantages, because then you dont have to finish and polish stuff. You just learn or try what interests you.
Lastly, side project is really not that dissimilar then working on the job - meaning job+side amounts to 10 hours a day work basically. It means your hourly effectivity will go down exactly like all studies on crunch predicts. You will get tired and slow.
It's a fairly clear signal of skill quality and attitude.
Reading open source commits/PRs and issue trackers tells you quite a lot about a developer which you can't see without some sort of a test (often not even then).
>it's biased against people who don't have the time
Surely any career that requires a high level of skill and practice honing that skill is biased against people who don't have the time?
What's special about open source?