After the data the collection, it is time to start the technical realization of the project. This means that all data have to be implemented in an Internet application. This procedure is following four basic goals:
- building models
- developing the user interface
- adopting the software
- testing, implementation of improvements, retesting
Why do we need models? What is the model?
As already indicated, we need many data, on the one hand dealing with barriers, POIs and references/hints, on the other hand addresses are important to locate certain infrastructures/facilities. Additionally we need a sidewalk network, again with information detailing the characteristics for a proper description of the route. In order to use ways2see not only for the city of Graz but also for different other cities, we need to create models. By applying these models, data can be adapted automatically or at least partly automatically and we do not have to redevelop adaptation procedures for each new city.
The user interface
The user interface is comparably simple and this is with intent. If the menu is undemanding, the program is easy to handle. This is important, as it ensures that the user interface is suitable for screen readers. However, it is not simple to create a simple user interface! The technicians of our partner company SynerGIS paid a lot of attention to the fact that the most important screen readers, especially JAWS and NVDA, as well as Internet-Explorer and other important browsers or compatible with ways2see.
Adopting the software
The most important challenge of programming was the fact that ways2see has features, which are not included in the software yet. This means that parts of the software had to be newly developed and/or reprogrammed. This sounds simple but parts of the program codes are not accessible for us and we had to make detours. The Internet platform ways2see is an expert program, which includes the computation of routes and the calculation of circumcircles.
Testing, retesting and testing again
After having designed a working user interface with all necessary functions, we have still not reached the end of our project. Only by extensive testing, we are able to recognize and repair/adapt mistakes. In a first step, the project team is performing test phases; later on, tests follow by a small group of interested users in a living lab. After a test phase with a larger community and before the project is finalized and published, a limited publicity is performing additional tests. These tests will be finished in December 2017.