GUI Testing

Gui testing, why it is so important?

The answer for that question we can find looking at the developer point of view, his developing schedule and software optimization desire, And sometimes even small changes causing the user interface to act differently, or even not to react at all.

To test a GUI, you need a reliable method of finding components of interest, clicking buttons, selecting cells in a table, dragging things about. In fact you need a ‘Bot’ that will activate all The User Interfaces, Buttons, Windows, etc. like human user does.

The GUI test need a full description of the expected behavior of the User Interface controls, and its process is similar to any developing process.

Gui test also had to make sure that the GUI stands on the customer requirements as well.

GUI testing cannot be avoided by designing a very simple GUI with less activities as less as possible, because before you will know it, it will be a part of a bigger algorithm or process.

GUI Testing framework must be ready for future changes, sometimes even with in the same product development process.

There is a need that the testing framework would not break just because few more buttons were added to the GUI layout.

One solution for that is using components with naming capabilities, naming them according to their purpose.

Many GUI testing systems have the ability to read this name and record it.

One the synchronizing is done by the object name, the risk for updated/maintain the testing script because few controls were added is drop-down to minimum.

The object naming option should be well discussed with the developers and bring it to the highest knowledge as possible. Don’t ever use the auto-naming mechanism, or no name option, or same name option as its parent object, this will cause the automation to be much less reliable and much more maintenance need.

Should the Testing frameworks check or expect for object hierarchy availability, in many cases it will raise the testing system reliability but on the other hand, the system will be more sensitive for changes in the GUI hierarchy and structure. Using the Parent window title is useful as a first-pass discriminator in multi-window application.

Component class, is not sufficient to unique the object, since application sometimes using the same class for many controls. Some of the GUI Testing tools have Recording events capabilities, meaning they have a hookup mode, which record all objects and even during the testers operation, this saves a lot of time when creating or updating the test by making the writing process much easier thus faster.

Not all recorders support all events generated by the Application under test, but almost all support the Pointer motion and clicks, Button press and Key strokes. Other event necessary to run the script are: waiting for specific window, opening window and closing window.

The easiest way to write a script is to record a series of actions for later replay, adding checkpoints or assertions along the way to verify correctness. Eventually, you might want to add heuristics that strip out meaningless events, but keeping track of a few basic events is sufficient for good functionality. Some GUI testing tools has a broaden specification and can handle work and operation against DB, API, HTML, Mail, etc.

See the Testing Tools section for onsite existing tools for GUI testing. ?