Frankenstein Tests
summary
Claims that autogenerated Playwright tests become unreadable Frankenstein code by mixing JavaScript, Playwright code generator locators and no domain language and demonstrates a real failing example that uses getByPlaceholder, redundant clicks and no backend check. Shows a fix path: move locators to a module, rename functions to domain verbs like agreeToTerms and submitRegistration, extract constants such as validEmail, add an API assertion that the saved entry includes the submitted data, and use adapters or page object interfaces to decouple tool language from domain intent.
AI-written summary · about
chapters (5)
- 0:00Introduction to Frankenstein tests
- 1:53Problems with generated tests
- 4:22Refactoring tests for readability
- 9:10Page object model approach
- 10:34Solutions and conclusion
related talks
51mAdvanced Playwright Techniques for Flawless Testing
29mPlaying your tests wright
28mFrom generic to specific test automation architecture with Playwright
21mEveryone Can Easily Write Tests
31mStop Guessing A11y: Auto-Generate Playwright Tests from Your GraphQL Schema
20mTesting Web Applications with Playwright
20mBeyond Code Coverage: Functionality Testing with Playwright MCP
35mState of end-to-end testing for modern web applications