Testing should never be optional in the software development process. Unfortunately, there are still programmers that think that they can develop software without testing.
Software construction implies knowledge, experience, talent, intellectual capacity and even some art. In other words, it is a very difficult task, and there is still a long way to go before that changes for the better.
It is true that we have surpassed, decades ago, the human capacity for apprehension and memorization; which necessarily implies that failures and errors are inevitable if we try to avoid them with only our human capacities.
Testing is not optional; software without testing is a bomb waiting to explode. For this reason, testing is essential, since it allow us to guarantee that applications meet the expected functionalities, helping us to find those errors or defects that have not yet been discovered; reducing the cost of development, the cost of ownership for users; and build customer trust by avoiding annoying regression bugs.
Software testing is one of the most important and fundamental activities in the development of a project, since it enables the processes, work methods and tools necessary to guarantee the quality of any software development. However, today we continue to encounter problems when it comes to understanding software testing as an investment, even in the most traditional projects where we usually assume that testing activities are normalized.
In order to be able to detect any error in time and guarantee that the product complies with all the established premises, testing must exist in all phases of a software project: from taking customer requirements, through follow-up meetings, to the delivery of the final product. What’s more, a project that lacks this process in all its phases ends up generating a higher economic cost and a greater effort during the testing phase.
In most cases, during project planning, the QA team has a few days established for tests, which are usually a few days before launching it. In case of development delays, the testing phase is usually the most affected, and in some cases is even eliminated at all due to lack of time. It is often considered that software testing is an expensive and at the same time expendable phase, when what it really achieves is to ensure software quality; therefore, we should see it as a guaranteed investment. This is because detecting errors in the final phase, their subsequent corrections and the testing of these corrections generate greater effort in hours and greater economic investment, than having planned the existence of the QA team throughout the development life cycle from the beginning, thus being able to prevent errors in advance. The key is to prevent first, not detect last.
A good functional and technical understanding of the business by the QA team guarantees, in turn, what type of tests will be necessary to apply in each testing phase and how to apply them. Here is where automated tests come into play. It is proven that testing a functionality usually takes twice time than developing it. Likewise, the exploratory tests only manage to detect a percentage of errors, also influenced by the short time available to carry them out and the number of people available to carry them out.
Automated tests achieve the elimination of errors arising from tests carried out manually. In addition, for each software update, regression tests must be carried out to guarantee that the product continues to function correctly and that none of the modifications has produced one or more collateral errors. Regression tests produce an extra effort in hours, as well as a greater economic expense, which makes automation necessary, thus saving project time and improving reliability. To do this, it is necessary to analyze which test cases are likely to be automated, always prioritizing those that are most critical and those that are likely to generate uncontrolled errors. This is done by professional software testing companies, who are experts of these kinds of tests.
For example, in an application, managing a long form and sending data afterwards, or an information search screen with its corresponding filter, will always generate more errors than accessing one or more screens from a menu. Therefore, all those screens that interact with the backend or services will be the highest priority to be automated. On the other hand, through a cycle of continuous integration we will be able to minimize the time and effort in regression tests, speed up product development and updates, reduce the TTM (Time to Market), while achieving an increase in test coverage, minimizing execution errors.
In short, it is necessary to be aware of the importance of software testing and the inclusion of the quality team in all phases of product development, without this being considered an economic loss in project planning. Undoubtedly, it is much better to detect errors at the right time, which gives enough time to correct them, than to detect them before the delivery of the product or in an important step in production.
