Internal
Much of this information is internal documentation. Care has not been made to make links work and make the documentation conprehensible for outsiders.

New Interaction Methods for Complex Graphical Tools

The purpose of this project is to develop and experiment with new interaction techniques for tools offering a complex graphical user interface. These techniques can be classified into three broad categories:

  1. improving the representation of the objects to be manipulated, e.g., by the use of layers, rich graphical rendering and multi-scale representations;
  2. improving the interaction between the user and the system with new direct manipulation techniques (e.g., mark-based input, marking menus, toolglasses, two-handed interaction) and better feedback;
  3. improving the modelling of the user interface by creating a more object-oriented description of the interface and the interaction with concepts such as layers, styles, tools and constraints.

Layers are used in graphics software such as Adobe Photoshop to structure the drawing. The concept can be applied and developed for a wide variety of applications to help users manipulate complex data by structuring it into layers that can be independently manipulated. To help manage the complexity of the display, effects such as texture mapping, lighting and shadows, usually found in 3D interfaces, will be applied to 2D interfaces. Multiscale representations can also be used to control the amount of information to be displayed according to the scale at which it is presented (this is also known as ZUI or Zoomable User Interface).

Finally, the interaction can be simplified and made more powerful by “reifying” concepts and commands into objects. For examples, the style sheets commonly found in text processing systems can be generalised to any application where rendering (or other) attributes may be applied to different objects. For controlling the layout of objects, graphical objects can be turned into first-class objects that can be visualised and manipulated directly by the end-user. For example, objects can be aligned by “sticking” them to an alignment stick that can be manipulated directly by the user.

A number of these techniques have already been experimented with small to medium-sized prototypes. However, they have never been used together in a complex, “real-size” system. A major goal of the project is to design and implement a software layer (or toolkit) which will make it easy for developers to implement these interaction techniques and to create new ones.

As a test bed for the experiments, the project will design and implement parts of a new user interface for Design/CPN which is a tool for editing, simulation and state space analysis of Coloured Petri Nets (also known as CP-nets or CPNs). The use of Design/CPN is adequate because it involves manipulation of complex graphical objects for many different purposes:

  1. CP-nets are specified by means of annotated graphs consisting of nodes, arcs, regions and a variety of different textual elements.
  2. The hierarchical relationship between the different modules in a CP-net is shown by means of a hierarchy graph, which also plays the role of being an interactive project browser.
  3. During an interactive simulation the CPN token game is displayed directly on the CP-nets by means of different graphical objects which are added to the CP-nets.
  4. Other simulation results are displayed by means of business graphs, message sequence charts and different kinds of graphical animations.
  5. Some of the results of state space analysis is shown by means of graphs.

The Design/CPN tool is developed and maintained by the CPN group at the University of Aarhus. The tool is used by 500 different organisations in 40 different countries, including 100 commercial enterprises. Moreover, the tool is used in two different courses at the University of Aarhus. Hence, there will be good chances to test the new user interface and get feedback for further improvements.

One of the main contributors to the basic ideas behind the new interface techniques will be Michel Beudouin-Lafon who is visiting the HCI group at the University of Aarhus as a guest professor for a period of two years. Michel Beudouin-Lafon has worked with graphical user interfaces in nearly 15 years and he is internationally very well-known and highly respected for his innovative work in this area. As indicated above, the work will involve three different groups at the University of Aarhus: the CPN group, the BETA group and the HCI group. These groups are already collaborating in others projects and they are all part of the DEVISE Centre for Experimental System Development. Via the involvement of the HCI group, the project will collaborate with the Center for Human-Machine Interaction, funded by the Danish National Research Foundation (Grundforskningsfonden).

Scientific interest

The project will develop and experiment with new interaction techniques for tools offering a complex graphical user interface. It will also develop a framework and associated software toolkit to support the development of such user interfaces. This will advance the state-of-the-art both in graphical interaction techniques and software architectures for interactive systems, with applications to a wide range of domains including CASE tools, office productivity tools, CAD tools, etc. The results from this work will be published in high-quality IT conferences and journals.

The project also contributes to the following general objectives for the Center for Human-Machine Interaction:

  1. Develop an integrated methodology that can be applied in design of new interface and system concepts which increase cooperating actors' possibility for controlling dynamic and safety critical work with heterogeneous information. New concepts should support flexible adaptation to dynamic work situations with urgent demands for high quality and safety.
  2. Demonstrate new interface concepts through prototypes that apply new input-output modalities (3D interfaces, Virtual Reality, WWW, intelligent agents, intelligent instrumentation and equipment, etc.) and evaluate these new interface modalities empirically in work practice and by means of experimental data from laboratory tests. New methods for using new

Business interest

The primary interest for Hewlett-Packard is the development of an improved version of Design/CP. This tool is used in all the projects of the existing HP-CPN centre, and the quality of it is vital for the achievement of the goals for the centre. However, the project will also allow Hewlett-Packard to achieve first-hand knowledge of the new, innovative interaction techniques to be developed and experimented with.