What to expect from my first Selenium Conference?

Tuesday, October 15, 2019

A question I asked myself after receiving the invitation from my manager for the SeleniumConf 2019 in London. What is there to learn about a tool that I’ve been using on a daily base for the past 5 years? Well, not so much about the tool itself, apparently, but that is not going to prevent me from going to Selenium Conf as often as I can!

Selenenium Conf, day 1

An interesing introduction talk by Simon Steward

First of all, White October Events and all the sponsors did a marvelous job setting up this conference. Not ever were the guests lacking anything. Coffee, lunch, snacks, culture, swag (stuff we all get), and of course inspiring talks; it was all taken care off. Stepping into the introduction talk, State of the Union, by Simon Steward. An interesting roundup on Selenium history and Simons involvement in the project. The presentation was building up to big news, which I hoped to be the release of a stable version Selenium 4. But unfortunately (and understandably) that remains in alpha just a bit longer, until it is done. The big news was that Simon will release Selenium 4, but that will be Simons last release of Selenium as project lead. While Simon creates Selenium with many other great committers, I couldn’t help but feel a bit anxious that such a key player is going to step down.

Adding accessibility into Unit Tests

After a short break and Simon Stewards “Office hours”, a small timeslot offering attendees to discuss previously completed talks with the speakers, I attended “Adding accessibility into Unit Tests: improving audit results and our experiences” by Elizabeth Simister and Heidi Jungel. They explained the laws on accessibility for specific countries, and how they implemented (if I recall correctly) the Axe accessibility testing framework. For someone who never had to perform any accessibility testing it was an interesting introduction to this subject. While it differs not so much from other unit testing, their framework has some very nice reporting functionality, which, per error, explains why an error occurs and how this can affect accessibility for users. Like an attribute not being provided for an input field, which can result in text-to-speech engines not being able to identify the field properly.

The automation breakup

The talk of Mark Winteringham about “The automation breakup: saying goodbye to full stack tests with task analysis” was very interesting as well. It is about stepping away from full stack UI driven tests, and breaking these tests up into multiple tests addressing each layer of the stack. So a e2e login test would result in a UI test using Selenium to verify the presence of the correct fields in the front end, a REST test would be used to request a login token, a unit test would be used to verify the correctness of the token etc. Very interesting approach, which results in way better identification of errors. Reports will no longer say “Login fails”, but can be way more precise, like “Invalid tokens are provided by broker”.

Time for Valori's talk: 40 countries, 35 brands, 1 test automation framework

After that we quickly had to claim our front row seats at my colleagues Michel Lalmohamed and Wout de Jong’s talk: “40 countries, 35 brands, 1 test automation framework”. The reason why we were there in the first place. And they did great on their explanation how they managed to test up to 40 websites in numerous languages, all with their own layout and specifications, using only 1 testframework. All their struggles on maintainability, runtime and overhead were addressed, accompanied by clear examples and explanations. And it payed off; after they were done, there was a que of people who wanted to talk to Michel and Wout about their talk. And after their talk we decided that it was enough for the day and grabbed some drinks to celebrate their success.

Selenenium Conf, day 2

The second day, Richard Bradshaw kicked off with a more in-depth explanation of his test philosophy, and how his career evolved over the years, together with his approach to test automation. Mark had introduced us to some of this approach on day 1, but the in depth explanation of Richard was truly inspiring.

Information market with various stands

There were various stands located in the hallway of the Queen Elizabeth 2 conference centre. One that really caught my attention was TestProject; an automation platform which uses Selenium and Appium. Some nice IDE-esque tools are present and the community can create their own plugins and share them amongst each other. It has dashboards for maintaining multiple projects/cycles, and it is free. While it sounds too good to be true, they promised it would remain free and that made it a real promising first encounter with TestProject.

So, did I learn anything new about Selenium itself?

Not really, but SeleniumConf 2019 was a great way to gain new insights on automated testing, tooling and test strategies. Thanks Valori for the opportunity to be part of this valuable experience and I can surely recommend all automation specialists to be there (again) next episode, in India.