Rubin Aga
Writing

Lexicon

Terms used in the posts, each explained in a line or two. Inside a post, they open as a card when you hover over them.

API contract test concept

A test that calls the real backend and checks that it still answers in the shape the frontend expects. If the backend changes a field, the contract test fails before the app does.

Contract test, by Martin Fowler

Code coverage metric

The share of the code that runs at least once while the tests run. High coverage doesn’t prove the tests are good, but low coverage means some code is never checked at all.

Code coverage on Wikipedia

Decision model concept

A model that answers a typed question instead of writing text: yes or no, one of a few choices, or a score, each with a probability. Examples include TypeSafe’s Jev and OpenAI’s Decisions API.

Jev on AI/ML API

DeepL tool

A machine translation service with an API, known for natural-sounding translations in European languages.

deepl.com

Docker tool

Runs software in containers, so a whole backend can be started on any machine with one command, the same way every time.

docker.com

Edge function concept

A small piece of server code that runs on demand, close to the user. In Kapmo, edge functions write challenges, translate them and check photos.

Supabase Edge Functions

Expo framework

The tooling around React Native that handles builds, updates and native modules for you. Kapmo is built on it.

expo.dev

Fixture concept

Setup that a test gets handed before it starts, like a browser page that’s already logged in, so each test doesn’t repeat it.

Playwright fixtures

i18next tool

A library for translating apps. Each language gets a file of keys and their text, and the app picks the right one at runtime.

i18next.com

Landscape product

Canonical’s tool for managing fleets of Ubuntu machines: updates, packages and security across many computers at once. Its web frontend is open source.

landscape-ui on GitHub

MSW tool

Mock Service Worker. It answers the app’s network requests with fake data, so tests don’t need a real server. More in [Testing a large React app].

mswjs.io

Playwright tool

A tool that drives a real browser through an app the way a person would, clicking and typing.

playwright.dev

React Native framework

A framework for writing an app once in React and shipping it to both iOS and Android.

reactnative.dev

React Testing Library tool

A library for testing React components the way a person uses them: find a button by its label, click it, check what appears. Tests don’t depend on how the component is built inside.

testing-library.com

Sentry tool

Collects crashes and errors from apps in production, with the stack trace and the device they happened on.

sentry.io

Supabase platform

A hosted backend built on Postgres: database, sign-in, file storage and edge functions in one place. Kapmo runs entirely on it.

supabase.com

TICS tool

TIOBE’s code-quality tool. It scores a codebase on test coverage, complexity and coding standards, and Canonical tracks its projects on a TICS dashboard.

TIOBE TICS

TQI metric

TIOBE Quality Indicator. The overall score from TICS, combining test coverage, complexity, compiler warnings, coding standards, duplication, dependencies and security into one percentage.

TIOBE TQI

Vision model concept

A language model that can also read images, so you can give it a photo and ask questions about it.

Multimodal models on Wikipedia

Vitest tool

A test runner for JavaScript and TypeScript, built on Vite.

vitest.dev