GUI testing: from developing scripts to creating AI-enabled agents

TESTAR is a tool that implements a scriptless approach for completely automated test generation for event-based Systems Under Test (SUT). Once the tool has sufficient information about the characteristics of the states of the SUT and what actions or events the SUT expects in a specific state, then the tool can test the SUT fully automatically, without the use of programmed scripts. This is due to the agents that implement various action selection mechanisms and test oracles. The underlying principles are very simple: generate test sequences of (state,action)-pairs by starting up the System Under Test (SUT) in its initial state and continuously select an action to bring the SUT into another state. The action selection characterises the most basic problem of intelligent systems: what to do next.

The difficult part is optimising the action selection to find faults and recognising a faulty state when it is found. Faulty states are not restricted to errors in functionality. Violations of other quality characteristics, like accessibility or security, could be also be detected by inspecting the state.

TESTAR has been used in various projects and has shown to be successful a testing. This totally shifts the paradigm of event-based testing; from developing scripts to developing intelligent AI-enabled agents.

Key Takeaways:

  1. How scriptless testing works.
  2. How AI is used to make scriptless testing more intelligent.
  3. What is means to shift the paradigm from developing scripts to developing intelligent AI-enabled agents



<< previous next >>