5 Reasons Why Open Source Test Automation is declining
Appvance.ai CEO on modern testing in 2018
I recently predicted that open source test automation tools will continue to decline. While it won’t ever completely go away, further evidence of its decline is all around. Here are the top 5 reasons.
- Selenium and other open source tools/languages are simply outdated for modern web architectures. This has been true for a long time, but is becoming too painful to ignore. Selenium was great with HTML 1, but HTML 5 really is beyond its design limits. And when you add modern AngularJS and ReactJS libraries, coders spend more time figuring out how to interact with those elements than the framework is worth. Let’s face it, Selenium is 14 years old and showing its age. Further, Selenium 3 was not the upgrade everyone wanted.
- Enterprises thought they would save money, but in the end their real cost is their people. For instance, we have seen test teams with five people dedicated to supporting each open source tool. The cost of that team far outweighs the cost of buying a commercial system that comes with vendor support. Of course, we know this hidden cost to be true since Red Hat proved it was cheaper to buy their support of Linux than to support it internally. In the end, nothing is really free, which is why open-source has proven to be very expensive indeed for large test teams.
- Great coders are busy getting paid, not hanging around to spend time on an open source project. Thus, the best people are busy in real jobs, given a shortage of quality coders worldwide. This was not the case in 1998, when open source authors gained notoriety for the free source code they provided. Today, most open source tools have less than five actual contributors. Without a source of income, their contributions start to run thin.
- Testing has become very challenging to do right. If one wants to automate unit testing, functional and regression, performance and load, compatibility and security, that is a lot of tools (or one Appvance IQ system). But Appvance IQ is about 4 million lines of code and took approximately 250,000 engineering-hours to develop. This level of technology is well outside the scope of a couple of authors of an open source tool. So, it is no surprise that testers are ten to 1,000 times more productive on Appvance IQ than on a hodge-podge of open source automation tools.
- Open source community support is failing major enterprises. Sometimes there is great support; other times none. Test engineers are too busy to troll support groups and provide answers. There are far more unanswered questions in some groups than answered ones, and in the meantime, the test engineer must deliver results by tomorrow. When the release is counting on you and there is a question, you don’t want to “hope” that someone shows up to answer it online. You want a contractual obligation for a vendor to respond within the hour to address your need. In some ways, a great commercial offering is software with totally amazing service. At Appvance, we consider ourselves a support-first organization, meaning that our only job is to make sure you can complete your testing on time. Even if we need to dedicate senior engineers to write tests for you, we will get you there on time.
If your team is ready for an order of magnitude change, now is a great time to get a live demo on the calendar. Request yours today.
– Kevin Surace, Appvance.ai CEO