We're bad at web integration testing

on YouTube ↗04m

" summary

Claims that web integration testing mainly fails because common tools check the wrong thing: selenium is slow, old and tied to Java; phantom.js verifies only a specific WebKit and JS engine so it misses real-user browsers. Proposes embedding an abstract interaction description inside delivered web components so automation is component-driven, tests stay domain-aware and data-independent. Reports a prototype tool implementing this architecture and confirms it is used in production for a bit over a year.

AI-written summary · about

" related talks

• 29mState of Web Integration TestingMatti Schneider · Wildcard Conference 2013
• 53mModern Web Testing: Going Beyond SeleniumDmitry Vinnik · GeeCON 2019
• 38mModern Web Testing: Going Beyond SeleniumDmitry Vinnik · Voxxed Days Frontend Bucharest 2019
• 55mModern Web Testing: Going Beyond SeleniumDmitry Vinnik · NDC Minnesota 2019
• 30mModern Web Testing: Going Beyond SeleniumDmitry Vinnik · DjangoCon US 2021
• 19mWorking Well: The Future of Web TestingTrent Willis · dotJS 2017
• 36mModern Web Testing: Going Beyond SeleniumDmitry Vinnik · TestCon Vilnius 2018
• 32mAll about web application testingAnisha Narang · FOSSASIA 2015 DevOps Track