When the test cases are ready, you can execute them. You can use our Harmony runner that can be downloaded prior you execute your first test.
You can execute one test case, all the test cases for a feature, and all the test cases for a project. In this way, if the application is ready, then you can immediately try your test and fix it, if necessary. To do this it is essential that your test cases be independent of each other.
You can use the test cases of other features by which complex test cases can be made quickly.
Our experiences are that concrete test steps should be made from the model when the application is ready and should also be executed immediately after creation. If a test step fails you can compare the test step with the actual screen and immediately realise what’s the problem. Creating and executing the tests are very fast and you can automate the necessary test cases in-sprint.
Below is an example of executing test cases for our “Pizza” application. Here one test failed since the total price is 39.5 instead of 31.