The exponential curve behind open source backlogs
7 hours ago
- #Queue Theory
- #Development Workflow
- #Open Source
- A developer describes a year-long effort to merge a small feature into Jellyfin web, highlighting common open-source backlog issues.
- The backlog problem is widespread: CPython has over 2,200 open PRs, Vue.js faced overwhelming issues, and 60% of maintainers have considered quitting.
- Analysis using queuing theory shows that as resource utilization nears 100%, wait times grow exponentially, with the average PR cycle time at 6.7 months for Jellyfin.
- A death spiral occurs where large PRs pile up, making reviews slower and encouraging contributors to bundle more changes, further increasing review time.
- Slow feedback wastes time; a study found contributors abandon PRs due to obstacles and hurdles during review, not rejection.
- Suggested solutions include capping PR sizes, gating quality with automation, limiting work in progress, prioritizing by value, setting a review cadence, building a reviewer tier, and requiring proposals for features.
- The core issue is not having one maintainer but inefficient workflow management, where time is wasted on low-quality PRs without flow controls.