DevOps Online exclusively reveals insights from the Interim Project Manager GDPR at Reed Business Information, Steve Watson, about how the scope of software testing has changed since he began to test
Steve has been in testing for many years, working his way from Tester, to Senior Lead Manager, to now currently being the Interim Project Manager for GDPR at Reed Business, which is an interesting piece of legislation, as he gets to use his management skills in a different way – stakeholder and task management, organising a security review of applications, managing the backlog of remediation tasks, identifying third party vendors to be contacted, and working with colleagues – not just in tech, but also in the customer services, sales, marketing and conferences teams.
When he first began to test, everything was mainframe-based, manually tested, and there were no tools! He also had no training, and there was no certification to help him learn his job.
Watson revealed: “I began to test before the internet, so, of course, there was no interconnectivity, and no need to be so security aware as there were fewer ways for people to access the bank’s computer network I previously worked for.
Waterfall to agile
“Now, not only have we had e-commerce for a number of years, we have the ‘Internet of Things’, and all the challenges it brings with interconnected devices. The challenges of privacy and security are prevalent in ways we could not have dreamt of back in the late 1980’s, and early 1990’s.
“We also transitioned from waterfall to more agile methodologies, changing the way in which testers are involved in technical work, and the very nature of our roles (shift left). Finally, there is tooling – many testers write code to test developers code, which didn’t happen when I first began to test.”
Nowadays, AI is very much a talking point, and the creation of artificially intelligent machines appears to be a huge turning point for us humans. A lot of what testers do requires code, and most of what we produce is used by human beings.
“AI will change the scope of software testing, as fewer humans will have to take actions, because of more machine-to-machine types of devices; machines will make the decisions that humans would have done. We may care less about usability and how things look on a screen, as the end-user is not a person, in which case our focus will be much more on data and analytics,” continued Watson.
According to Watson, test environments always used to be a pain to manage – mainly due to not being kept up-to-date with the correct data or code changes, and there is always a cost element to requesting more environments, so often environment clashes would ensue.
“Now, with the advent of cloud computing, we can spin up a test environment when needed, and tear it down once finished. The data we need can be deployed, although with the advent of GDPR, we have to be careful about obfuscating any personal data that we may have in pre-live environments,” added Watson.
“In terms of tools, Selenium Webdriver has almost become the defacto standard for web application testing, and you can configure it to work with other tools such as Specflow or Gauge to add in business case definitions, and then code in whichever language you choose. We have used this for a number of years, adding SoapUI, Postman and other tools for API testing.”
He also noted that it’s hard to move from tool-to-tool when you have created a large set of tests, as it’s not ideal to be spending time reinventing the wheel, so testers now look at tooling if they are working on something brand new, or if they find that an existing test is not fit for purpose. We take a pragmatic approach to tooling needs, taking into account skills, time and budget.
Written by Leah Alger