For any software enterprise, reliable, effective testing is the cornerstone to fail-safe product delivery to the customer. ARCAD software has developed powerful automation tools, such as its i Unit and Verifier that can greatly automate and improve both your unit testing and functional test processes. These tools, however–like any test tool–are reliant upon one essential dependency: good test data.
Whether they be a developer running a fairly simple unit test, or QA personnel executing a complex regression test, testers need consistent data for successful execution. On the other hand, poorly-managed data sets can shift or become corrupted during the test process itself, resulting in false-positives. As often is the case, the job falls on the developer to recreate the test failure–an impossible task, considering the data has already changed. One can see how preservation of the integrity of test data leads to more reliable results, quicker troubleshooting, less false-positives, and valuable time saved. And the key to good test data is consistency.
Unit testing involves a developer running tests against their isolated code component, and the results of such tests are typically valuable only to the programmer in validating the functionality of the code itself. Data set requirements are also typically less demanding than those of functional regression tests, and ideally, test data can also be well controlled. With many legacy IBM i systems however this is unfortunately not the case, as the necessary test data can indeed have other data dependencies chained deeply into obscure application business functions. In this case, data must be updated each time prior to test or special test scripts may be written to take these conditions into account. ARCAD i Unit has been designed specifically to automate the unit test process and also automatically searches for parameters and data types (including complex data structures). Additionally, it can mock all kinds of test input, from data to files and other key components.
Functional regression tests, on the other hand, are executed against various business functions of an application, such as UI, database, or security features, and check a user input against an expected output. Functional tests are scripted and reused many times, and ARCAD Verifier can be extremely helpful in managing and automating this process. Ideal functional tests work off a dedicated data set, otherwise–since tests are running in the actual application environment–other processes not related to the test can also corrupt the data. Verifier has the added benefit of using a test repository to reinitialize data prior to each test, ensuring that only the changes of interest to the software are being tested. ARCAD’s DOT Anonymizer can also help in protecting high-value sensitive data that is often necessary for the test.
ARCAD’s various test tools can save your company time and money, as well as ensure against many of the pitfalls of testing. But regardless of the tool you used, reliable testing starts and ends with the same foundational principle: consistent, well-managed data.