n8n

A canary release is a deployment strategy where a new version is gradually rolled out to a small subset of users before reaching the full audience. Automated monitoring compares error rates, latency, and key metrics between the canary and the stable version. If the canary performs well, traffic is incrementally shifted; if anomalies are detected, the release is automatically rolled back, minimizing the blast radius of potential issues.

Atomic Design is a methodology by Brad Frost for creating design systems by breaking interfaces into five hierarchical levels: atoms (basic HTML elements), molecules (simple component groups), organisms (complex UI sections), templates (page-level layouts), and pages (specific instances with real content). This approach provides a consistent mental model for organizing component libraries in React, React Native, or Flutter projects and ensures systematic scalability of the UI.

In Flutter, the widget tree is the hierarchical structure of nested widgets that defines the entire UI of an application. Everything in Flutter is a widget — from layout structures like `Column` and `Row` to styling wrappers like `Padding` and `Theme` to interactive elements like `GestureDetector`. Flutter maintains three parallel trees (widget, element, and render object) that work together to efficiently rebuild only the parts of the UI that actually changed when state updates.

BEM (Block Element Modifier) is a CSS naming convention that structures class names as `.block__element--modifier` to create clear, predictable relationships between HTML and CSS. For example, `.card__title--highlighted` identifies a highlighted title element within a card block. While modern tooling like CSS Modules and utility frameworks have reduced BEM's necessity, it remains widely used in large codebases and design systems where explicit naming conventions aid team coordination.

A context window is the maximum amount of text (measured in tokens) that an LLM can process in a single interaction, encompassing both the input prompt and the generated output. Larger context windows allow models to handle longer documents, maintain extended conversations, and reason over more information at once. Context window sizes have grown rapidly — from 4K tokens in early GPT models to over 1M tokens in current models like Claude.

Focus management is the practice of controlling which element receives keyboard focus in a web application, especially during dynamic content changes. When a modal opens, focus should move into it; when it closes, focus should return to the trigger element. Proper focus management is essential for keyboard and screen reader users, and involves techniques like focus trapping, roving tabindex, and programmatic focus with `element.focus()`.