EN FR
EN FR


Section: Software

TouchStone

Participants : Caroline Appert [correspondant] , Michel Beaudouin-Lafon, Wendy Mackay.

TouchStone [8] is a platform for designing, running and analyzing the results of controlled experiments (Figure 5 ). While it focuses on experiments comparing interaction techniques, it can be used in a wide variety of contexts.

Figure 5. The architecture of the Touchstone platform
IMG/touchstone.png

With the Touchstone design platform, a user specifies the factors and the measures of the experiment, the blocking and counterbalancing of trials, and assess the time it will take to run the experiment. Multiple designs can be explored in parallel to assess the various trade-offs. The output of the design platform is an XML file that can be used as input for the run platform.

The Touchstone run platform provides a framework to implement and run an experiment and to collect experimental data. It uses a flexible plug-in architecture to manage a variety of input devices and interaction techniques. The runs of the experiment are controlled by an XML script that can be produced by the design platform.

The analysis platform currently consists of data analysis tools such as JMP, R or Excel. Log data produced by the run platform can be directly loaded into any of these tools. In a future version, analysis sketches will be derived from the experimental design to assist with the analysis.

Touchstone has been used heavily at INSITU over the past three years for the many experiments that we design and run. It has also been used for teaching for the first time in 2011. Students used it to design various experiments during tutorial classes in Master 2 Interaction (“Introduction to HCI” module).

Touchstone is available at http://code.google.com/p/touchstone-platforms/ under a BSD License.

  • ACM: H.5.2 [User Interfaces]: Graphical user interfaces (GUI)

  • Software benefit: See W. Mackay, C. Appert, M. Beaudouin-Lafon, O. Chapuis, Y. Du, JD. Fekete and Y. Guiard (2007) TouchStone: Exploratory Design of Experiments. In Proceedings of ACM CHI 2007 Conference on Human Factors and Computing Systems. ACM, pages 1425-1434.

  • OS/Middleware: Mac OS X, Linux, Windows

  • Required library or software: Java virtual machine

  • Programming language: Java