Performing the first test
A major milestone in a project is being able to test at all. Getting a feature working for the first time means that all the code is in place and working in the test environment. It has been successfully moved from the development system and works elsewhere, be that in a full test environment, a test harness, your local machine, or a containerized environment (see Chapter 5, Black-Box Functional Testing, for a discussion of test systems). Getting the feature running opens up a vast array of possible tests.
For instance, if you are testing the signup page of a website, does that page load and accept input? If so, then there are many follow-on test types you can perform. If not, let the developer know that you can’t even start yet.
Carrying out that first test requires many development tasks to be completed. Another easy way to waste time is by testing a feature that hasn’t been enabled yet. You can also solve this with better communication with the development team.
The specification will say what a feature should do (see Chapter 2, Writing Great Feature Specifications), but testers need another level of detail, which is how the feature is configured. Before testing begins, ensure you understand the following:
- What the minimum version requirements for all relevant parts of the system are
- What the necessary configuration is
- How to check that a feature is working
Version requirements are clear enough – you have to be running the version with the new code before you can test. However, sometimes, it is far from obvious which parts of a system are dependent on each other. Or, while the feature is implemented on the 5.3 branch, is it in build 5.3.8 or 5.3.9? A feature may be delivered piecemeal, in which case, exactly which functionality is in each build?
For your very first test, only try the most basic functionality. Does the new screen load, are API calls accepted, or is the new option available? Be clear on which versions you need before spending your time on the first case. Which feature flag do you need to enable for this feature? Which setting needs to be updated, and which file is that in? Again, the challenge is to get all the necessary details from the development team to avoid wasting time looking for them.
If you have all the requirements in place but have found several blocking issues, check with the development team that this feature is ready for testing. The developer should have done enough testing on their system to be confident that it will work for others, but that isn’t always the case. If there are repeated problems, get the developer to recheck their code.
Finally, how can you tell if the feature is working? Sometimes, features are obviously customer visible – is the new web page present, or does the new option appear? Sometimes, however, it’s hard to tell if the new feature is enabled, especially during code refactoring or subtle performance changes. Which log lines should you look out for? What statistics will indicate this change is being used?
Those details won’t be in the feature specification; again, this is an extra level of detail that the test team requires to check on the behavior or even the existence of a feature in a particular build of code.
From a project point of view, getting the first test running is on the critical path. Being able to start testing delays everything else, so make sure you complete that early. Don’t spend ages getting everything in place, for instance, finishing six other projects so that a large team of testers is ready to descend on a feature… only to find that it’s not working, and you can’t test it at all. Check the feature early to make sure the functionality is basically in place, and quickly bounce it back to the developer if they need to make any changes. Once that first test has passed, you can leave it for a while until you can test it properly. But make sure it’s ready first.
Once you have the feature happily running, then the real work begins. Where to start with exploratory testing will be described in the next section.
