Robb Owen built a Svelte/Netlify quiz app to learn Latvian noun cases, then found he’d already learned them by writing the code.
Key Takeaways
Shipping is not the only valid success metric: learning a Go API, wrangling GIS data in Postgres, or building a broken animation that later became a portfolio site are all concrete wins.
The quiz app used Svelte 3, Netlify serverless functions, static JSON for noun data, and regex-based stem-stripping to generate and check 84 possible Latvian noun endings.
Hustle-culture pressure and recruiter GitHub audits inflate the perceived cost of abandonment; the post argues the cost is usually near zero when learning happened.
Treating side-projects as throwaway prototypes removes shipping pressure and turns them into low-stakes experiment pads.
Retrospectives on abandoned projects reveal what was actually learned, same as a work post-mortem but for personal code.
Hacker News Comment Review
Commenters largely agreed that “abandoned” is the wrong frame: several noted projects get paused, not killed, and can be picked up years later when skill or context changes.
A recurring thread pushed further than the article: on a side-project you can intentionally ignore best practices, reinvent wheels, and make decisions based on curiosity alone, which is itself a feature, not a bug.
One commenter proposed a concrete ritual the article stops short of: write an “end of life wrap-up” document when stopping, covering what was built, learned, and why you’re stopping, so closure feels deliberate rather than like failure.
Notable Comments
@mjd: argues abandoned projects often get finished decades later once skill catches up, and middle age is evidence for that claim.
@roland35: used Claude and ChatGPT to kill a years-long side-project idea, found the AI verdict “liberating” and redirected to more successful work.
