Continuous integration (CI) has emerged as one of the most efficient ways to develop code. But testing has not always been a major part of the CI conversation.
In some respects, that’s not surprising. Traditionally, CI has been all about speeding up the coding, building, and release process. Instead of having each programmer write code separately, integrate it manually, and then wait until the next daily or weekly build to see if the changes broke anything, CI lets developers code and compile on a virtually continuous basis. It also means developers and admins can work together seamlessly, since the programming and build processes are always in sync.
Unfortunately, the quality assurance team does not necessarily reap the same benefits. While CI assures that your app keeps building successfully as the code is continually updated, it doesn’t automatically test how new builds behave within different types of environments. An otherwise well-run CI operation might require app testing to be done separately, on a non-continuous basis, instead of building it into the rest of the process.
This poses real problems for an organization. Unless you add automated testing to your CI mix, you could end up with an app that users can download and install properly, but which suffers from critical usability issues in certain browsers or operating systems. Arguably, an app that installs successfully but frustrates users due to lack of testing is worse than one that doesn’t install at all. (more…)