Ask HN: How do I find “up-and-coming” open source projects to work on?
2 comments
I think the best way to find a project is to scratch your own itch. What kinds of problems do you face that an existing tool could solve, if it just had that one feature? That is how I ended up contributing to maintaining Seleniarm for RaspberryPi and ARM machines.
Also, don't worry about asking questions. Most maintainers are more than happy to help new contributors. Many projects need all the help they can get.
Also, don't worry about asking questions. Most maintainers are more than happy to help new contributors. Many projects need all the help they can get.
I see, that's a cool way of looking at it! Sadly I don't have many "itches" right now, mostly cuz I've done a lot of "write it 100% from scratch for fun" type projects for the past year or two. But hey, that's a nice mindset, and ig I'll just keep my eyes open until something catches my attention.
Yeah I'm OK with asking questions, but there definitely are projects with enough domain knowledge that it would take upwards of 30 minutes to an hour to fully explain everything... And that's kinda where my comfort limit of asking questions ends :P
Like there's this mildly embarrassing occurrence where I saw an open issue and decided to tackle it. Made it to the point where a maintainer asked for multiple live zulip discussions, but it became pretty clear that I was missing lots of domain knowledge, so after a while these discussions kinda tailed off and IDK if the PR would be considered dead or not. If so, I probably just wasted 4-5 hours total of that guy's time lol. Like that's the type of situation I'd like to avoid :p
(cherry on top is I missed an obvious-in-hindsight mapping of the problem to a graph theory problem, so my PR was 1000 lines of code when it probably could've been shorter. Then again, IDK if there's that many graph automorphism libraries I can just pull off the shelf anyways, so maybe there wasn't much potential for simplification)
Yeah I'm OK with asking questions, but there definitely are projects with enough domain knowledge that it would take upwards of 30 minutes to an hour to fully explain everything... And that's kinda where my comfort limit of asking questions ends :P
Like there's this mildly embarrassing occurrence where I saw an open issue and decided to tackle it. Made it to the point where a maintainer asked for multiple live zulip discussions, but it became pretty clear that I was missing lots of domain knowledge, so after a while these discussions kinda tailed off and IDK if the PR would be considered dead or not. If so, I probably just wasted 4-5 hours total of that guy's time lol. Like that's the type of situation I'd like to avoid :p
(cherry on top is I missed an obvious-in-hindsight mapping of the problem to a graph theory problem, so my PR was 1000 lines of code when it probably could've been shorter. Then again, IDK if there's that many graph automorphism libraries I can just pull off the shelf anyways, so maybe there wasn't much potential for simplification)
I've spent some time working on these large projects (various libraries and tools, a brief foray into the rust compiler), but it often feels like the only work I'm able to do amounts to just small ad-hoc bugfixes and refractoring, which while I understand its importance, doesn't really "feel" satisfying.
Ideally, I'd love to find a cool project that's far enough along that the owner/maintainers are accepting contributions, but still early enough that the whole project can be kept in your head pretty easily, and still requires significant development. Does anyone know a good place to find these projects? (perhaps requests for contributions or something?)
Of course, one could always say, "why not just make that project yourself", but sadly coming up with an original, impactful idea is quite difficult. I've played around with everything from making parser generators to 3d games, but they're just solo projects that always end up as pale imitations of other projects - fun to make, but not impactful at all.