Development of an ed... Theses on offer UML-basierte Beschre...

Student thesis

Formulation of criteria and indicators for the design of robust Human-Machine-Interfaces

(Original Title:Erarbeitung von Kriterien und Indikatoren für das Design robuster Mensch-Maschine-Schnittstellen)



The goal of this work was to find criteria for the design of a good user interface. In the first part of this work it was tried to identify properties for well-designed user interfaces and to define criteria how to check whether an interface is designed reflecting these properties. Afterwards these criteria should be checked by a self-implemented application by performing an experiment with several test persons. Every user interface should have the following seven properties having been standardized in ISO norm 9241, part 10:

  1. Suitability for the task
  2. self descriptiveness
  3. controllability
  4. conformity with user expectations
  5. error tolerance
  6. suitability for individualisation
  7. suitability for learning.
These seven properties should apply to textual user interfaces as well as to graphical user interfaces. To judge whether an interface meets these properties, criteria are needed. Additionally graphical user interfaces have several special features that were dealt with in the following paragraph. These specialities are:
  1. the use of colours
  2. the use of fonts
  3. the design of formulas
  4. the design of menus
  5. the design of radio-buttons and check-boxes.
For these elements, too, criteria have been found. In the practical part an application was implemented and with this application an experiment was carried out with several test persons. This was done to find out how much the found criteria of part one would impact the usability of the application. The application consisted of an address book but the focus was not on setting up complex functionality but instead on the possibility to start the application with different configurations. The application was built up in a way that allowed to start the application with different user interfaces in order to test different criteria of part one of the work. For the test the users had to deal with some given exercises by using one of the versions of the application. Their way of solving the problems was protocolled. By evaluating these protocols it was analysed using which version of the application the users could do their work best. So it could be concluded which criteria, when being broken, impacted the usability of the application worst. The result of the experiment was that breaking some criteria had really bad consequences for the usability of the user interface while breaking other criteria didnít impact the usability of the interface very much.

Author: Claudia Schieber

Tutor: Dipl.-Math. Jens Palluch