The development of AVoS follows a user-centered-design approach. This means that users from the target group are included in the development process. The development is cyclic with the following four stages: (1) a goal is set, (2) it is implemented, (3) it is tested with users, and (4) it is evaluated. Then the goal is re-assessed and the cycle begins anew.
Before the actual implementation an exploratory studies is conducted to assess the concept. The following gives an idea of how the design of the auditory feedback patterns has been found in collaboration with visually impaired people, describing three full cycles of usability testing. Firstly, the concept (0.0) was evaluated. To this end, humans correct their spelling errors orally, in a situation where they do not see what they have typed. Particular feedback patterns are recorded. Secondly, the recorded feedback (1.0) was evaluated. The test moderator played the feedback where appropriate and helped the test person in case it was insuﬃcient. Thirdly, the feedback that worked best was automatised within a prototype (2.0) and evaluated again. The picture below illustrates the methodology used.
It becomes clear that an exploratory study at this stage allows a greater degree of freedom compared to testing with an existing prototype. The test moderator is able to change their approach in order to find out which strategies work best. The goal is for the system to respond kust like a real language tutor would. It is hence useful to know how a real language tutor would react. Efficient teaching strategies are then more and more automatised. A solution is implemented only when a concept has been proven successful with the users. At this point, the prototype is evaluated again to find out which features are still missing.