---
title: "0002. Rank correctness first, then ergonomics, efficiency, and completeness in the core"
description: "Accepted, 2026-10-20"
canonical_url: "https://docs.galleon.thec1oud.uk/docs/decisions/0002-rank-correct-ergonomic-efficient-complete"
markdown_url: "https://docs.galleon.thec1oud.uk/docs/decisions/0002-rank-correct-ergonomic-efficient-complete.md"
last_updated: "2018-10-20"
x_farming_labs_generated_preamble: true
---

# 0002. Rank correctness first, then ergonomics, efficiency, and completeness in the core
URL: /docs/decisions/0002-rank-correct-ergonomic-efficient-complete
LLM index: /llms.txt
Description: Accepted, 2026-10-20

# 0002. Rank correctness first, then ergonomics, efficiency, and completeness in the core

**Status:** Accepted · **Date:** 2026-10-20 · [Source on GitHub](https://github.com/thec1oud/galleon/blob/dev/docs/adr/0002-rank-correct-ergonomic-efficient-complete.md)

## Context

Galleon aims to be correct, complete, ergonomic and efficient. Those goals
collide constantly: typed extractors against a quick `c.Get("user")`,
refusing to boot on a route with no auth policy against a fast prototype,
shipping every feature against a core a small team can maintain. Without a
ranking, every such collision is argued from scratch.

## Decision

Galleon ranks its goals **correct > ergonomic > efficient > complete in the
core**.

Efficiency has a hard budget: no ergonomic feature may add per-request
reflection or allocation on the hot path unless the user opts in. Each
component is to document its hot-path cost, and benchmarks in CI are to
catch regressions.

Completeness is delivered through batteries and seams, not a bigger core.

## Consequences

- A correctness bug in a framework is copied into every app built on it, so
  correctness wins every tie.
- "Efficient" becomes reviewable: a change either adds hot-path cost or it
  doesn't.
- Most of the conflict between ergonomics and efficiency is resolved at boot
  (the DI graph, validation plans), or is to be resolved by generated code
  (generated decoders), not per request.
- The core stays small; a feature that isn't needed for correctness goes in a
  battery.

## Alternatives considered

- **Efficiency above ergonomics.** For most apps the database and network
  dominate latency. Fiber shows the cost: speed bought with compatibility and
  safety.
- **Completeness in the core.** Completeness is the easiest of the four to
  add later, and a large core is what sank Buffalo.

## Sources

- [Part 4: decision D5](https://github.com/thec1oud/galleon/blob/0897d74/docs/design/04-desires-debate.md#decisions-so-far)
- [Part 4: ergonomics vs efficiency, in Round 2](https://github.com/thec1oud/galleon/blob/0897d74/docs/design/04-desires-debate.md#ergonomics-vs-efficiency-d5-ergonomics-with-an-efficiency-budget)
- [Part 4: Pushback 7](https://github.com/thec1oud/galleon/blob/0897d74/docs/design/04-desires-debate.md#pushback-7-correct-complete-ergonomic-will-conflict-so-rank-them)
- [Part 5: principle 1](https://github.com/thec1oud/galleon/blob/0897d74/docs/design/05-architecture.md#0-principles)

## Sitemap

See the full [sitemap](/sitemap.md) for all pages.
Docs-scoped sitemap: [/docs/sitemap.md](/docs/sitemap.md).
Well-known sitemap: [/.well-known/sitemap.md](/.well-known/sitemap.md).