DEV Community

Cover image for An Effective Approach to Selecting Test Cases for Automation
QA Madness
QA Madness

Posted on • Edited on

An Effective Approach to Selecting Test Cases for Automation

Software development companies are in hot pursuit of automation whether it’s set up in-house or as outsourced testing. End-users are demanding higher quality apps with more advanced, intuitive features, which has spurred the evolution in software testing methodologies. Manual software testing is by no means obsolete, but it is not sufficient as a single approach to delivering high-end, bug-free products. It is time-consuming, overly-dependent on human resources, and it doesn’t ensure reliable and consistent outcomes.

However, automating every aspect of software testing is not the ultimate solution nor the end goal. Managing too many test cases can be counterproductive and impractical that can cause the testing cycle to slow down with delayed feedback to developers. All too often, teams scramble to automate everything only to run into a mountain of problems.

Test Automation Best Practices

There are several factors to consider when identifying initial test candidates for automation testing. There are different testing techniques, and each has their place in the testing process. Some techniques are best-suited to manual testing while others are excellent contenders for automation.

Not every test case is suited for QA automation and this particularly true for subjective testing that requires exploration or improvisation. Automated tests only test what they are programmed to test.

Automation Criteria

Successfully implementing test automation starts with careful thought and planning resulting in tests that are trustworthy, repeatable, informative, and focused. To get the most benefit from automated test cases, define which tests apply to the following criteria:

  • Tests that are frequently run for multiple builds.
  • Tests for critical functionality and high-risk conditions.
  • Tests that need to be run on different hardware and software platforms.
  • Tests that are tedious and slow when run manually.
  • Tests involving multiple data sets in a data driven framework.
  • Tests that deliver an exact pass or fail result.

These parameters will help the QA team select the right tests for automation, and help set standards and prioritization across the team to ensure timeous and quality releases.

Automation Testing Types

Generally, there are three types of testing methods that take priority for automation with Unit Testing being the first, and then followed by Integration Testing and Functional Testing. This can later be expanded on with automating regression testing, system testing, as well as user acceptance testing.

After identifying the test cases that can be automated, the next step is developing the automation test. It is important to start with creating small, logical tests as large, complex tests can be cumbersome to edit and debug. The automated test environment must be manageable and coherent.

Once the initial automated tests are added to the Continuous Integration process, you can grow your automation library over time to group smaller test cases into larger executable tests.

Unit Tests

Unit testing focuses on small parts of the code logic and should be top priority for automation because they are easy to develop and debug, and are highly reusable. Unit tests should make up the biggest portion of automated tests.

Unit tests are an important element as they check the code functionality without having any dependencies. When the code is changed and contains errors, the unit test will pick this up before it goes into the test environment and QA team. A popular approach is for developers to write the unit tests before writing the code, known as test-driven development (TDD).

By applying TDD, it helps the developer consider all use cases first. Automated unit testing should be executed every time a piece of code is changed or added.

Integration Tests

Integration testing should be the next step for automation. It involves testing modules and/or interfaces that are required to work with the software. Forms of integration tests include synchronicity between mobile and web apps, use of different devices, cross-browser testing, launching modules, and interacting with email clients.

Today, the majority of software applications are cloud-based and need to seamlessly run in a web browser, and there are several popular web browsers such as Safari, Google Chrome, Firefox, and Microsoft Edge.

Automated cloud integration tests are more efficient than manually testing in each different environment. Start off with basic tests on the most popular platforms, such as loading the application in each browser to verify the user interface renders correctly.

The objective of automating integration tests is to quickly detect any error with the application interacting with dependent interfaces and modules. Having stable test data available and managing timing issues (for example with page loading) are key to executing successful automated integration tests.

Functional Testing

Functional testing focuses on the end-user’s journey and on the application providing the correct output based on a set of actions and it involves interacting with the user interface. Automated functional tests can be implemented the moment a working and stable UI is available.

Automate functional tests that mimic the most common workflows that users take like creating a new account, resetting a password, logging into the application, or going through a checkout and purchase process. These automated tests can run every time a new build is compiled without any human intervention making it faster to test the most common or critical functions and features.

At a later stage, automated functional tests can include tasks and actions that are not obvious or a divergence of expected user behavior to test errors are handled appropriately without the application crashing.

UI Testing

By automating User Interface tests, you can minimize a lot of effort validating areas of the UI and it significantly increases the test coverage. Through correct configuration and implementation, it is possible to create test scripts for different user scenarios and user inputs that simulate the user’s journey through the application. Automated UI testing is quicker and more accurate than manually clicking through the interface to verify UI controls and data outputs.

API Testing

API testing focuses on the business logic and does not involve testing the UI or source code. Automating these test cases is a rapid approach to detecting connection issues between systems, missing or duplicated functionality, security, and multi-threading problems. It also helps to determine if there are performance flaws as the response time between API should typically be very high. Automated API tests can be run every time a build is compiled that will also ensure proper handling of values and errors, and validate the response data is structured correctly.

Best Tools for Automation QA Engineers

A test automation framework is a vital part of successfully adopting a true Agile and DevOps environment and minimizes bottlenecks in the product development lifecycle while helping to produce quality software. However, implementing automation requires a good toolset that can be used by both developers and testers.

There are many open-source and licensed tools available for API, mobile, web, and desktop applications, and the choice of automation tools will depend on the business’s skillset, requirements, and budget.

Some of the top open-source test automation tools include:

  • Selenium is the number one choice for automating web browser testing across several platforms, supports a large number of programming languages, and can be integrated with a lot of other tools. It has a steep learning curve but offers extensible use in the testing process.
  • Selendroid is the mobile Selenium version native and hybrid Android mobile app testing on different devices.
  • Appium is an automated testing tool for iOS, Android, and Windows mobile applications. It can automate apps written in any language using any testing framework.
  • Calabash supports the development and execution of mobile QA for iOS and Android apps. It works with Cucumber, which allows business users to write test cases in simple English using Gherkin syntax and doesn’t require knowledge of programming. It also supports languages such as Ruby and Java.

In Summary

The goal of test automation is not intended to replace a QA engineer’s responsibilities but rather to support and improve the whole testing effort. Effective automated test cases allow for quicker feedback to development, quicker detection and resolution of issues, and frees up time for testers to focus on exploratory testing.

Automated testing is a piece of the testing puzzle that works in parallel with manual testing as a combined approach. Successfully managing both is crucial to ensuring the success of your software product.

Top comments (0)