Software Testing: Reactive vs Proactive - GenRocket

Software Testing: Reactive vs Proactive - GenRocket

Software Testing: Reactive vs Proactive

by HYCEL TAYLOR on MARCH 31, 2014

In some testing environments, there is a belief that using production data to test your software will ensure that you will find bugs and therefore create the right tests both in quality and quantity.

There are too many things wrong with this way of testing.  If one is waiting for bugs and edge cases to be found in production in order to fix and prevent bugs in the code base, then it points to major weaknesses in the testing environment.  Some of these weaknesses include:

• Inadequate testing prior to production deployments which leads to the high probability that many bugs will appear in production yielding a bad user experience and possibly the corruption of production data.
• Few, if any, unit tests for testing small blocks of business logic.
• Few, if any, integration tests to check the results of larger interacting units of code.
• Few if any load tests to check the code under load.
• Little to no functional tests, manual or automated, to validate the user interface is stable and bug free.  
• Test data that is written around production data is production dependent which means, as soon as the production code changes (and it will), all of the tests dependent on static production data conditions and formats will fail or at minimum become obsolete and no longer used.