Selenium Community Live - Episode 8

Selenium Community Live - Episode 8

Episode 8 of Selenium Community Live with Dorothy Graham took place on July 30th, 2025. During this remarkable session, Graham a legend with 50+ years in software testing, shared incredible insights from her journey through the evolution of test automation. Her perspective reveals timeless truths that modern teams often overlook.

Meet the Speakers:

  1. Dorothy Graham

Let’s dive into session notes.

From Mainframes to Smartphones: The Incredible Journey

Graham’s automation story began in 1970 at Bell Labs, working on a UNIVAC 1108 mainframe that cost $1.6 million (equivalent to $15 million today). The specifications, 1.3 megahertz, half a megabyte of RAM, and 100 megabytes of storage. Her iPhone today has 12,000 times more storage and costs 20,000 times less. Back then, you would write code on paper, punch it onto cards, and get maybe one turnaround per day. A single typo meant starting over tomorrow. This dramatic shift in just 50 years raises the question: what will the next 50 years bring?

The Evolution of Testing Tools: From Commercial to Open Source

Graham witnessed the dramatic transformation of testing tools over decades. The first commercial tool, AutoTester, appeared in 1985, followed by an explosion of tools in the 80s and 90s. The CAST (Computer Aided Software Testing) report eventually documented 103 different tools, yet when she recently checked, only three were still alive. What happened to all those tools? They disappeared, reminding us that tools come and go, but principles endure. Enter Selenium in 2004, a game-changer as an open-source tool that broke the expensive commercial tool monopoly. Graham congratulated the community: “Selenium has been around for over 20 years. That’s really good.” Its longevity stems from community support, continuous evolution, and freedom from licensing costs. However, Graham warns: “The fact that you don’t have to pay purchase or licensing costs doesn’t mean it doesn’t need investment.” Free tools still require proper architecture, training, and skilled implementation. She stresses, this as a reminder to people in management making decisions.

What Test Automation Shouldn’t Look Like

Graham’s most compelling insight involves what not to do in automation. She shares Steven Norman’s brilliant analogy: Imagine recording your drive to work, then pressing play the next day. You would reverse into traffic that wasn’t there yesterday, stop at green lights because they were red yesterday, and run red lights because they turned green yesterday. She mentions, Testing isn’t passive observation, it’s active investigation. Graham also emphasizes: “Testing is something you do. It’s not passive. It’s active.” This is why capture-replay approaches fundamentally misunderstand what good testing requires.

EuroSTAR Survey Insights: The Reality of Test Automation in 2023

In preparation for her keynote at the EuroSTAR conference in Vienna, Graham conducted a comprehensive survey of 200 automation practitioners that revealed surprising insights:

Encouraging findings:

• 80% use open-source tools.

• 90% test at system level, 72% at API level

• 70% had formal testing training

Concerning discoveries:

• 25% had zero training in their daily tools

• 38% would need to rewrite 95%+ of tests when changing tools

• Most automated tests mirror manual test structure (which is wrong)

• Top problem isn’t technical, it’s unrealistic management expectations

if you are interested to learn more about the survey, visit link

The Forgotten Secrets of Good Automation

Drawing from her decades as a practitioner, coach, and consultant, Graham shares what she believes are the most overlooked principles in modern test automation.

  1. Effectiveness Before Efficiency Graham’s first hard-learned lesson: the biggest mistake is automating poor-quality tests just to make them faster.

• Wrong approach: Poor manual tests → automation → fast, poor automated tests

• Right approach: Improve test effectiveness first → then automate selectively

She emphasizes: “You will get much better results if you think first about how can we improve our testing. You are better off having better testing than automating testing.”

  1. Proper Architecture Matters Through years of consulting, Graham identified that good automation requires two critical abstraction levels: Technical level: Modular, reusable scripts where one manual test becomes multiple automated scripts, and one script serves hundreds of tests. As she notes: “If you have a thousand manual tests, they might be implemented by only 50 scripts together with data files.” Business level: Domain-specific keywords like “create_new_policy” instead of raw code, enabling skilled testers to write tests without programming. This makes automation accessible to “people who are non-technical who often are the best testers.”

A Special Thank You to Our Community

This incredible session featured engaging questions from community members whose thoughtful inquiries sparked valuable discussions about management expectations, career paths, AI’s impact, and the future of testing skills. Their participation exemplified the collaborative spirit that makes the Selenium community so vibrant.

We extend our heartfelt gratitude to Dorothy Graham for sharing her wealth of knowledge and to all community members who joined Episode 8.

This session marks the end of Season 1 of Selenium Community Live, a season that began by celebrating two decades of Selenium and concluded with timeless wisdom from one of testing’s most respected pioneers.

As we prepare for Season 2, we invite you to stay connected with the Selenium community.

Watch the Recording

Couldn’t join us live? Watch the entire episode here - 📹 Recording Link: Watch the Event Recording on YouTube

Stay tuned as we bring the next!

Subscribe here to the Selenium HQ Official YouTube Channel.