Visual Modelling of and on Tangible User Interfaces

The purpose of this thesis is to investigate the use and benefit of visual modelling languages (VMLs) in modelling on tangible user interfaces (TUIs), and modelling TUI applications. Three main research questions are asked:
- Is it possible and practical to model an application or process using a TUI?
- Which General-purpose VML (GPVML) performs best in modelling a VML scenario for use on TUI?
- Is it realistic to use a GPVML to model complex TUI applications?

To answer the first research question, a Business Process Model and Notation, version 2 (BPMN2), ideation scenario is used with a tangible widget toolkit prototype on a TUI table to evaluate the performance and obtain feedback from test candidates. The study showed that it is possible to model a process using a TUI and the test candidate feedback did not lead to the conclusion that the modelling was cumbersome or impractical.

To find a suitable GPVML, the thesis explores and evaluates the current state of the art in VMLs for describing general problems. After gathering different VMLs, the thesis compares three languages using a simple scenario: the visual object constraint language (VOCL), augmented constraint diagrams (ACD), and the visual contract language (VCL). A weighted evaluation based on multiple quality criteria led to the conclusion that VCL is best suited to model TUI applications, answering the second research question.

The thesis answers the third research question by using VCL to model a more complex and complete scenario of an ideation process, which is based on using a BPMN2 on a TUI. This is done to assess VCL's suitability to more complex problems and its maturity. The study concludes that VCL is not yet mature enough to enable its general applicability in a wide variety of settings.

The three research questions were dressed with a hypothesis in mind: collaborative, novice friendly modelling environments are able to reduce the gap between stakeholders and software engineers during software projects, leading to a reduction of unrealistic expectations and an increase in the availability of domain knowledge. While the hypothesis is too broad to be
proven by this thesis, the research questions answered here give some insights into how to approach it.