Test Automation: Time to move away from Selenium Webdriver?
Back in the day, if you wanted to do automated UI testing you had to install Visual Test 6 (Showing my age here so be kind). That then died off and then you had to pay stupid money (i.e. ￡4 to ￡7k per license) for the likes of IBM Functional Tester (now Rational Functional Tester) or QuickTest Pro (now Micro Focus Unified Functional Testing) for your automated testing pleasure. Selenium Core came along in 2004, IDE in 2007 and Webdriver 2.0 came out of beta in 2011 to change the landscape as it was free.
Webdriver 2.0 was buggy at the time and had its foibles but it allowed the likes of me to develop cutting edge frameworks that could do more than QTP could ever do with none of the cost implications. Webdriver was awesome, it stabilised over time and although it could not do mobile devices, Dan Cuellar was doing his thing & introduced Appium in 2012 so all was good in the world.
The original plan was for Webdriver 3.0 to come out Christmas 2013 with the stated goal of "We aim for Selenium 3 to be a tool for user-focused automation of mobile and web apps”. Anyone who wants a trip down memory lane should visit https://seleniumhq.wordpress.com/2013/08/28/the-road-to-selenium-3
So where are we today now that the Webdriver API is almost a decade old?
The honest answer is although it remains a marvellous tool in the toolbox, it has stagnated quite badly. Version 3 was finally released in 2016 but mainly consisted of a massive refactoring of the architecture. Version 4 is currently in Alpha but the scope of changes are pretty basic and nothing to write home about. The last stable Java bindings is 3.141 & was published almost 2 years ago (14 Nov 2018). Don't get me wrong, I like Webdriver but it needs a lot more love than its been getting as of late. It's now living in a world where there is competition in the form of Puppeteer, Playwright & Cypress and in my opinion, webdriver is losing its mojo.
As an example, tools like WebdriverIO (which is one of my favourites) has moved away from using the Webdriver protocol under the hood and is using Puppeteer as its default instead. Cypress has likewise come on leaps and bounds in the last 12 months and is showing no signs of slowing down. We're soon going to be spoilt for choice so picking the right tool for the job long term is going to be important.
If you have an old webdriver framework (usually in Java) that has loads of test scripts that take ages to run and are flakey. Now might be a good time to evaluate the alternatives for your teams and plan for the future.
If you are still using the likes of UFT/QTP... I shall pray for your mortal soul :-)
If you've already looked at the alternatives & webdriver still does it for you and you feel none of the others meet your needs... all power to you.
Just remember that we live in an Agile world which means adapting and improving, continuously. What was the best tool 5 years ago may not mean its the best one for you now