dontreinvent.org  /  worth building  /  open source

The commons already solved it. Look first.

Before you write a line, search the shelf. Before you build, check what's already standing. Every unsolved problem is a space someone is probably already occupying, quietly, in the open.

Where to look

The commons is not one place. It is layered, overlapping, and often hidden beneath its own success. Start where friction points to.

  • 01. Code archives GitHub, GitLab, Gitea Search by problem, not by name. Read issue threads. Maintainers often explain why a feature does not exist.
  • 02. Package registries npm, PyPI, crates.io, Maven Central If it is a known problem shape, someone has probably published a tested answer. Check the last update, not the creation date.
  • 03. Language forums Discourse, Stack Exchange, mailing lists Look for the debates. Those conversations contain the tradeoffs that no README captures.
  • 04. Specification documents RFCs, standards, design docs The why behind the what. Why a choice was made shapes whether your problem is even a real gap.
  • 05. Open issues Unfixed problems in projects you trust An open issue is an open door. Closed issues are often already solved upstream; you just need to update.

What you find when you look

Looking first teaches you the shape of a problem before you solve it. You find:

The answer already exists

It is published, tested, and probably maintained by people who care about it more than you will. Use it, credit it, report bugs upstream.

The problem is different than you thought

The commons has seen more variations than you have. A conversation in an issue thread shows you the edge cases, the broken approaches, the real tradeoff you were not seeing from the inside.

The solution is half-built somewhere

Not quite what you need, but close enough to fork, wrap, or extend. A small bridge is often faster and safer than starting from scratch.

The problem is unsolved for a reason

Sometimes there is a silence in the commons because the problem has a bitter tradeoff. People tried. It did not work. That silence is worth hearing before you start.

What happens when you find nothing

Looking first is not about always finding an answer. It is about making sure you know what you are committing to when you start from blank.

If you have read the specifications, searched the registries, asked in the forums, and nothing moves, then building is not procrastination. It is the only move left. And when you do, you build knowing what you are not reinventing, which is a different thing entirely.

The commons does not need another half-informed wheel. It needs the wheel built by someone who has read enough to understand why the round shape was chosen.

Things the commons has already solved

A small, incomplete foundation of problems you almost certainly do not need to solve twice:

HTTP caching JSON parsing Cryptographic hashing Logging systems Rate limiting CORS handling Error recovery Database migrations Session management Authorization Date handling Compression formats Retry logic Configuration management Testing frameworks Code formatting

For each of these, start by looking. The answer is almost never zero, and almost never as good as the one you will find after you check.