Always be Closing
One of the biggest mistakes that an open-source project maintainer can do is to allow the backlog of their projects to grow to a size where it’s unmanageable. That something I’ve seen many times and it always happens in more or less the same fashion, and leads to more or less the same result.
You start a small personal project driven by your own interests and needs. The scope of the project is tiny at the time - just a handful of features, no plans for world domination. At some point the project gains some traction.1 Users (understandably) start reporting bugs that you never encountered and requesting new features you never envisioned. You’re excited that someone found your small hobby project useful and you are eager to oblige and expand the scope to keep the new users happy. If you’re extra lucky2 you start receiving lots of pull requests from contributors who are willing to help move the project forward. World domination is at hand!
Fast forward a few years into the future. Unless by some miracle you’ve got all the free time in the world or your project became your full-time job, you’re probably facing now hundreds of open tickets and dozens of open pull requests. No matter how hard you try you can never manage keep up with them. And you try really hard. You’re so busy these days that you can’t even keep up with the contributions you’re getting, let alone find some time to trim the backlog, label tickets accordingly, respond to everyone and handle a couple of tickets yourself.
And in those rare moments you find some time to peruse the backlog, so you can pick some tickets for yourself, you’re just getting depressed by its sheer size (e.g. 100+ open tickets) and you don’t know where to begin. You spend all of your time considering which ticket to start working on and then you’re out of time. No progress. Lots of frustration.
I happen to be the steward a few somewhat popular projects and I’m saying all of this from experience. I know that focus is the key to success in everything, but I’ve been finding it harder and harder to find my focus when it comes to OSS projects, as so much work has piled on my plate there that I constantly wonder where to begin. Time to make some changes.
Yesterday I’ve decided to enable a bot that automatically flags stale issues and pull requests on my major projects (e.g. RuboCop, CIDER, Projectile) after a certain period of inactivity and closes them later if they remain dormant. I always feel bad when I’m closing valid bug reports and interesting feature requests, but I think it’s much more important to be realistic about what’s actually feasible versus relying on wishful thinking and giving fake promises. If a ticket is not handled in a reasonable amount of time that usually means two things:
- Probably it’s not something super important.
- It’s unlikely that it will ever be handled (lack of bandwidth, the maintainers don’t feel it’s particularly important, it’s something super complicated, etc).
I’ve noticed that really important topics always tend to resurface and there’s always someone willing to tackle those. That’s why I’m not particularly worried about something major slipping between the cracks, when relying on an automated ticket clean-up strategy.
I hope that the outcome of this experiment is going to be a clean and focused roadmap for my projects resulting in better pace of development and happier users in the end of the day. Time will tell. At least I’m glad that I’ve finally tried to beat the vicious cycle I was trapped in, as you know what they say about people who keep doing the same things expecting different results…
Never forget that “less is more”. Keep closing!