Decision frameworks

Status: 🟩 COMPLETE Last updated: 2026-06-21 Plain-English tagline: “When to use X vs Y.” Short, opinionated guides for the recurring choices in modern webapp building — with a clear default and the conditions that change it.


What these are

A decision framework is a structured way to answer a recurring “which one?” question. It gives you:

  1. A short answer — the default for most cases
  2. The factors that matter — what would push you toward the alternative
  3. Specific “when to pick X / when to pick Y” sections
  4. What if you’ve already chosen — migration paths if you change your mind

The textbook entries cover what each option is. The decision frameworks cover which to reach for. Together they answer “what is it?” and “should I use it?”


The frameworks

ChoiceWhen it comes up
Server Component vs Client Component 🟩Every Next.js file you write
Merge vs rebase vs squash 🟩Every PR merge, every “I want to clean up my branch”
RLS vs server-side checks 🟩Every Supabase table that touches user data
Custom auth vs Supabase Auth 🟩Every new project’s auth decision
Free tier vs paid tier 🟩 🟦Vercel, Supabase, GitHub, every service with tiered pricing
TypeScript strict vs loose 🟩Every new TypeScript project’s tsconfig.json
Tailwind vs CSS-in-JS vs plain CSS 🟩Every new React/Next.js project’s styling architecture
When to write tests (and when not to) 🟩Every “should I unit-test this?” moment
When to refactor (and when to leave it alone) 🟩Every “this code bothers me” moment
Client-side fetch vs Server Actions 🟩Every Next.js mutation flow
Self-host vs SaaS 🟩Every infrastructure decision in the stack
Monorepo vs polyrepo 🟩Every “should I split this repo?” moment
Claude vs ChatGPT vs Gemini 🟩 🟦Choosing your primary AI assistant
Cursor vs Claude Code vs Windsurf 🟩 🟦Choosing your AI coding tool
AWS vs Azure vs GCP for AI 🟩 🟦Choosing a cloud for AI workloads
Western vs Chinese AI 🟩 🟦Why the encyclopedia recommends against Chinese AI tools
Paid AI subscriptions worth it? 🟩 🟦Free vs paid AI decision framework
Image vs video vs music AI 🟩 🟦Choosing the right creative AI medium
AI for small business 🟩 🟦Practical AI toolkit for Australian small business
AI for families 🟩 🟦AI at home; kids; ages; safety; what’s worth using
Local vs Cloud AI 🟩 🟦When to run AI on your own computer; hardware needs
AI for accessibility 🟩 🟦AI tools for disability; vision, hearing, mobility, cognitive
AI for content creators 🟩 🟦YouTube, podcasts, social media AI toolkit
AI for students 🟩 🟦Using AI to learn, not cheat; academic integrity
AI for writers 🟩 🟦Fiction, journalism, professional writing; ethics
AI for older adults 🟩 🟦Genuinely helpful tools for older Australians; scam awareness
AI for travel planning 🟩 🟦Real help vs marketing hype; what to verify
AI for mental wellness 🟩 🟦Honest framework; helps and risks; crisis support
AI for cooking 🟩 🟦Recipes, meal planning, kitchen help
AI for educators 🟩 🟦Teachers’ practical AI toolkit; student AI issues
AI for tradespeople 🟩 🟦Practical AI for trades businesses
AI for real estate 🟩 🟦Agents, investors, buyers, renters; Australian property context
AI for Australian mining 🟩 🟦World-leading AI deployment; autonomous trucks/trains/drilling
AI for Australian agriculture 🟩 🟦Precision agriculture; livestock; AgTech ecosystem
AI for non-profits 🟩 🟦Free/discounted tools; grant writing; ACNC context
AI for new immigrants to Australia 🟩 🟦Settlement; translation; navigating systems
AI for fitness 🟩 🟦Wearables; workout apps; nutrition; Australian fitness companies
AI for journalism 🟩 🟦News, ethics, MEAA, defamation, transcription, verification
AI for marketers 🟩 🟦Content, campaigns, brand voice, Australian Consumer Law
AI for HR professionals 🟩 🟦People & culture, recruitment AI, Fair Work, anti-discrimination
AI for hospitality 🟩 🟦Restaurants, hotels, tourism, multilingual, Australian context
AI for healthcare practitioners 🟩 🟦Unified clinical framework; AHPRA, TGA, indemnity
AI for accountants 🟩 🟦Tax, bookkeeping, Xero/MYOB AI, TASA, CPA/CAANZ
AI for lawyers 🟩 🟦Unified Australian legal practice; citation discipline; privilege
AI for government workers 🟩 🟦APS, state, local; security, privacy, Robodebt lessons
counsellors 🟩 🟦APS Code, AHPRA, ACA, ethics, sensitive practice
AI for retail 🟩 🟦E-commerce, physical retail, ACL compliance

This folder now covers the recurring choices that come up across the Next.js + Supabase + Vercel stack. Future entries land here only when genuinely new recurring choices surface in real work.


How to use these

Read once for the mental model. Most of the value is the structure for thinking, not the specific recommendation. Once you know what factors matter, you can apply them to new situations.

Re-read when you face the choice. The defaults are the right call for most cases, but the “when to pick X” lists catch the cases where the default is wrong.

Don’t be afraid to change your mind. Every framework has a “what if I’ve already chosen?” section. Migration paths exist; you’re not locked in.


What’s NOT a decision framework

Some choices look like they need a framework but really don’t:

  • “What language should I use for my Next.js project?” — TypeScript. (No serious tradeoff for the typical case.)
  • “Should I use Git?” — Yes. (See What is version control? 🟩.)
  • “Should I run npm run build before pushing?” — Yes. (The playbook quality bar, not a framework.)

A decision framework exists when the right answer genuinely depends on your situation. If the answer is the same for ~everyone, it’s a recommendation, not a framework.


See also

47 items under this folder.