With the pervasive waterfall methodology, software testing has always traditionally been a fairly settled affair - get to the end of your project cycles and get your testing team on it. In recent years this has all changed with trends toward automation and artificial intelligence.
2018 has already brought with it some promising indications for the way things are heading for software QA for the remainder of the year and the future beyond.
1. Fully Automated Regression Testing
Regression doesn’t ever really need to be manually tested. The fact that roughly 80% of companies still manually test for regression is a bit baffling to those who are up to speed with modern QA.
Regression tests by their very nature are hugely time consuming, laborious and require next to no creativity. This is especially the case in hugely complex systems, where it’s simply impossible for a single team to try to break a system using ‘exploratory’ methods.
All this points to a key trend we are seeing - that automation can, and is slowly being, adopted instead of manual testing for regression. Plus, the breadth and specificity of tools out there grows by the year, so keep an eye out for solutions that fit your industry and needs.
2. Artificially Intelligent Test Automation
Automation is all well and good, but the really effective test automation tools coming out in the near future are going to be the ones which can utilise artificial intelligence to aid with all aspects of the testing process.
AI can be used in software testing from the bottom up, not only predicting user patterns and points of high load within a system, but in establishing test cases and writing test scripts. Artificial intelligence will become especially important in test maintenance. Less brittle tests means less time spent maintaining and more time on development.
Smart analytics are also going to become front and centre in measuring product and team success as well, so watch that space.
3. Shortened Product Delivery Times
The mindset within the technology community is growing to be that you must analyse and update your project methodologies increasingly frequently to iterate your team to success. If you’re up with the times, you’ll know that release cycles are getting shorter and shorter, so therefore programmers are required to be even more responsive to changes in the industry. If there’s a more effective way of handling your projects you should be considering reading about it now.
In the ever faster moving world of technologies, if you haven’t developed the next ‘must have’ feature, then you’re going to be behind. With shorter release cycles grows the need for more testing, more automation and a better understanding of what various testing options can provide you.
Along the same theme as above, DevOPs are the set of practices aiming to reduce the time between development and Operations. No doubt you’ve already heard this term before, but it’s something that’s really taking off, and has been for the last few years.
Of course, making the transition from the developing floor to the operations room as quickly as possible in complex environments leaves room for errors, so new testing methodologies must be implemented to cope with DevOPS strategies.
One of the main tips we suggest for implementing DevOPs is daily full-system testing, made possible by automation - catching bugs as they are made, gives you the opportunity to rectify them then and there before they become seriously integrated into your code and costly to fix.
5. ‘Internet of Things’ Testing
The idea of the ‘Internet of Things’ has become all the rage the last few years. If you are reading this newsletter as an implementer of enterprise asset management, then no doubt you have seen or heard about IBM’s press materials about IoT and its relation with EAM.
Everything, including your assets, has brains these days. With that, systems are becoming so complex that they are growing to be unmanageable for a single team to fully test when changes are made.
Testing IoT systems comprehensively requires automation no matter what you do - it’s almost impossible to test the functionality of all aspects of such a maze-like network of different devices.
The theme of 2018 is looking like shorter release cycles and better test automation, so it's time to start thinking how your testing can be modified to suit your needs in an increasingly changing world.