SDFrame

290 Software Design Frameworks — a curated, bilingual knowledge base for engineers, architects, and AI agents.

sdframe.caldis.me | Inspired by PMFrame.works

Overview

SDFrame organizes 239 software design frameworks across 13 categories, spanning the full software lifecycle from design thinking to AI-native collaboration patterns. Each framework includes:

• Bilingual content (EN/ZH) with natural translations
• 5 actionable implementation steps with visual diagrams
• Multi-dimensional taxonomy (abstraction level, maturity, quality concerns)
• Provenance citations to authoritative sources
• Typed relationships to related frameworks (alternative, complement, extends, prerequisite)

Categories

#	Category	Frameworks	Description
1	Design Thinking	23	Mental models, design philosophies, problem framing
2	Architecture Decisions	23	Architectural patterns, trade-off analysis, system decomposition
3	Coding Practices	39	Code-level design, GoF patterns, implementation
4	Quality Engineering	25	Testing strategies, reliability, observability
5	Deployment & Operations	20	DevOps, SRE, infrastructure patterns
6	Evolution & Iteration	20	Refactoring, tech debt, migration, team scaling
7	AI Collaboration	25	Agent architecture, LLM design, human-AI workflows
8	Data Architecture	15	Storage, streaming, data modeling patterns
9	Security & Privacy	20	Threat modeling, zero trust, privacy by design
10	Distributed Systems	20	Consensus, consistency, fault tolerance
11	API Design & Integration	20	REST, GraphQL, gRPC, versioning
12	Team & Organization	20	Team topologies, engineering culture, inner source
13	Observability & DX	20	Tracing, SLOs, developer experience

Features

Knowledge Base

• 290 frameworks with 39 fields each — bilingual, authoritative, cross-referenced
• Multi-dimensional taxonomy — filter by abstraction level (Code/Component/System/Organization), maturity ring, quality concerns (ISO 25010)
• Typed relationships — 5 relation types: alternative, complement, extends, prerequisite, related
• Provenance — primary source citations and recommended reading per category

Pages

• Home — horizontal scroll card rows with lerp engine, staggered entrance, fingerprint radars
• Framework Detail — 11-section layout, scroll progress bar, section nav dots, entrance animations, similarity recommendations
• Category Landing — category overview with AI cross-reference and reading list
• Relationship Map — D3 force graph with label collision detection, typed edges, search, side panel
• Compare — searchable framework picker, radar chart overlay, deep links, diff highlighting
• Selector — 4-step wizard to find frameworks matching your situation
• Learning Paths — curated sequences to master software design step by step
• Insights — editorial data visualizations (category bars, dimension triptych, quality grid, heatmap, adopter sizing)
• Timeline — frameworks plotted by origin year, grouped by decade

Technical

• Dark mode — warm-tinted palette, auto-detect + manual toggle, 55+ CSS variables
• Two-tier data loading — listing stubs (603KB) upfront; detail data lazy-loaded per category (76% reduction)
• Search autocomplete — dropdown suggestions with keyboard navigation and text highlighting
• PWA — service worker, manifest, offline caching
• i18n — full language switching with browser auto-detection
• SEO — Open Graph meta tags, sitemap, per-page titles
• 54 E2E tests — smoke, visual/a11y, interaction state, data display, screenshots
• Mobile — bottom tab navigation, responsive layouts, touch scroll snap

Tech Stack

• Vite 6 + React 19 + TypeScript
• React Router 7 — client-side routing with lazy-loaded pages
• Recharts — radar charts, data visualization
• D3.js — force-directed relationship graph
• Playwright — E2E testing
• CSS Modules — Minimal Scholar aesthetic (serif headers, warm palette)
• GitHub Actions — CI/CD to GitHub Pages

Development

npm install
npm run dev        # Dev server
npm run build      # Production build
npm run test:e2e   # Run Playwright E2E tests

Data Pipeline

Framework data lives in data/frameworks/*.json (one file per category). A pre-build script generates lightweight listing stubs for fast initial load:

data/frameworks/*.json  →  scripts/generate-stubs.js  →  src/data/generated/stubs.json (upfront)
                        →  dynamic import() per category (on demand)

Project Structure

data/frameworks/         # 13 JSON files (239 frameworks)
scripts/                 # Build scripts (stubs, sitemap)
src/
  components/            # Reusable UI (cards, filters, modal, viz)
  data/                  # Loader (sync stubs + async detail), categories
  hooks/                 # useSearch, useFavorites, useFrameworkDetail
  i18n/                  # Internationalization (EN/ZH)
  pages/                 # Route pages (7 pages)
  styles/                # Global CSS and custom properties
e2e/                     # Playwright E2E tests
.harness/                # AI agent workflow (sprint planning, evaluation)

Adding Frameworks

1. Add entries to the appropriate data/frameworks/<category>.json
2. Run node scripts/generate-stubs.js to regenerate listing stubs
3. Verify: npm run build && npx playwright test

Adding a Language

1. Create src/i18n/locales/<lang>.ts implementing LocaleStrings
2. Add the locale to Locale type in src/i18n/types.ts
3. Register in src/i18n/context.tsx
4. Update LanguageSwitcher component
5. Add _<lang> suffix fields to framework data

License

MIT