Skip to content

comparisons.md

Latest commit

 

History

History
76 lines (54 loc) · 3.76 KB

comparisons.md

File metadata and controls

76 lines (54 loc) · 3.76 KB

How Facet Differs

Facet vs Other Testing Approaches

Approach Focus Facet's Difference
TDD Code correctness via unit tests Facet tracks what is covered, not how to write code
BDD Behavior via Given-When-Then Facet uses free-form natural language, not structured syntax
ATDD Acceptance criteria drive development Facet maps multiple perspectives to tests, not just acceptance
Traditional Coverage Lines/branches executed Facet measures requirement coverage, not code coverage

Key insight: TDD/BDD/ATDD are development methodologies. Facet is a coverage framework. Use them together.

Facet vs Cucumber (and Similar BDD Tools)

Cucumber, SpecFlow, Behave, and similar tools follow a specification-as-test architecture where .feature files written in Gherkin syntax are directly executed as tests. Facet takes a fundamentally different approach: specification-as-documentation.

Aspect Cucumber / BDD Tools Facet
Syntax Gherkin (Given-When-Then) Free-form Markdown
Specifications Are the tests (executable) Are documentation (linked to tests)
Step definitions Required for each step Not needed - tests are regular code
Learning curve Must learn Gherkin syntax Write in natural language
Perspectives Single specification file Multiple facets per feature
Test framework Cucumber runner Any framework (Jest, Vitest, Playwright, etc.)
Maintenance Step definitions can drift from specs Direct linking prevents drift

Architectural Difference

Cucumber Approach:

Feature file (Gherkin) → Step definitions (glue code) → Test runner (Cucumber)

Facet Approach:

Facet files (Markdown) ←→ Your tests (any runner)
         ↓
   Coverage report

Why This Matters

  1. No translation layer - Cucumber requires step definitions that translate Gherkin to code. These become a maintenance burden and can drift from specs. Facet links directly to existing tests.

  2. Use your existing tests - With Cucumber, you rewrite tests in Gherkin. With Facet, you add facet() calls to tests you already have.

  3. Multiple perspectives - A Cucumber feature file is one perspective. Facet lets business, compliance, UX, and technical teams each write their own facets for the same feature.

  4. Natural language freedom - Gherkin's structured syntax (Given/When/Then) forces a specific format. Facet accepts any prose—bullet points, paragraphs, tables—whatever communicates best.

  5. Framework agnostic - Cucumber requires the Cucumber test runner. Facet works with Jest, Vitest, Mocha, Playwright, or any testing framework.

When to Use Which

  • Cucumber: When you want executable specifications and your team has invested in the Gherkin ecosystem
  • Facet: When you want coverage tracking across multiple perspectives without changing how you write tests

Why Facet Works for AI-Driven Testing

AI coding assistants excel at generating tests but struggle with what to test. Facet solves this:

  • Natural language specs - AI understands requirements without parsing Gherkin
  • Multi-perspective facets - AI generates tests covering business, compliance, UX in one pass
  • Type-safe linking - AI can programmatically verify coverage completeness
  • Gap detection - AI identifies untested facets and generates missing tests
Human: "Generate tests for checkout"
         ↓
AI reads facets → Understands business rules, PCI compliance, UX requirements
         ↓
Generates comprehensive tests → Links to facets automatically

Facet bridges human intent and AI execution with traceable, verifiable coverage.