Discussion
I've switched many of my open-source projects to yearly release cycles, and honestly, it's great.
Turns out software development is much more enjoyable when you stop pretending every Tuesday is a historic event.
Instead of switching contexts 8 times a day, I now switch projects about once a month.
Speaking of releases, there's a fairly big one coming up that I've been quietly working on for a while now.
If you're into terminals, CLI tools and all things TUI, you might want to keep an eye on this space.
End of transmission, no screenshots, no details, no hints. I'm really good at this teasing business. Back to your regularly scheduled programming.
This probably isn't ideal for a brand-new project with a mountain of TODOs. But for mature projects, I've found that giving yourself more time between releases is incredibly valuable:
You stop treating every idea as an obligation and every bug report as an emergency.
Turns out a little distance between having an idea and implementing it can do wonders for software quality. Occasionally you will realize that doing nothing was actually the right call.
@fribbledom I think I understand why and fully support this. Would you still be willing to release a minor fix in between for a severe security problem, should one be reliably reported? Not saying you should, I am genuinely interested in how you feel about such a situation.
Yes, absolutely. These are usually fairly small fixes that, once identified, aren't particularly difficult to review.
This isn't meant to be a rigid rule. Priorities and schedules change. It's more about being deliberate with my time. I think both the projects and I benefit from fewer context switches and longer stretches of focused work.
People will just have to get used to the idea that 12 months of radio silence doesn't necessarily mean a project is dead.
@fribbledom excellent! One reason I am asking is because we are discussing standards to signal to users of an application or a library that the software is still maintained even if no releases are happening right now. In your case you could announce that for the next year or one and a half no new releases would be needed and the software could still be regarded as maintained. In my understanding such a signal would help you and the users of your software. Comments?
@fribbledom
I feel genuinely bad for developers that aren't old enough to have been involved with the waterfall model of development.
Most will never know what it's like.
@fribbledom yea a major release about every year and a half is my limit. At the same time I don't stop doing major releases even though they're are quite a lot of people using point releases because it's a mechanism to drive bug hunts.
@fribbledom I hate seeing open issues is my problem.
Ah well, once you reach four- or five-digit numbers, you stop caring.
@fribbledom That would give me major anxiety
@fribbledom so with "release" you also mean dependency updates (minor/patch versions)?
Because I can imagine that the amount of work increasing those after a year might be way bigger than doing small, incremental dependency upgrades.