Switching Between Tools in Complex Applications
Journal of Usability Studies, Volume 3, Issue 4, August 2008, pp. 173-188
Article Contents
Conclusion
Tasks that are the work of large software applications are typically complex. Neither the user nor the usability tester knows either the path the workflow will take or which tools will be brought to bear. Users' awareness of points where they change tools, and the choice of tools they make at those points, directly affects productivity and creativity. Training and the inculcation of best practices may well improve this awareness and decision making ability, but most users never get either. Improved design for making these choices and transitions, on the other hand, reaches every user every time a tool needs changing.
A model of user action made up of steps or segments that smoothly merge into an efficient operation as expertise increases falls short as a design goal for complex software because it omits the following:
- Switching from one tool to another at times that depend uniquely on task content and progress. Users must become or be made aware that the work is not advancing as well as it might be, and that a tool change is the right next move.
- Consideration and study of complex, "real" tasks wherein the user discovers or invents the workflow. The switch from one optimal tool to another at the appropriate time is a key problem in this discovery and invention of a workflow.
- Discovery and adoption by experts of better, more appropriate tools-simply tools that are newly available. As soon as the task is fresh, or in any way different, the expert's polished workflow, however impressively effective, is potentially ripe for improvement. Experts must cope with this possibility without losing their effectiveness.
We have only vague notions of how design may help users to use the best tools in the best order, and how to test proposed improvements. We need to replace these notions with studies and observations.
Acknowledgements
Andrew Wirtanen and Mike Ryan, who facilitated the testing, Amy Kidd, who helped me plan it, Jared Spool, who taught me how to design real tasks, Donna Cooper, the MATLAB LCT Usability team, Chauncey Wilson, the excellent peer reviewers, and others who grappled so supportively with ideas out of order, and, of course, the participants.