The project
In the late ’90s, one of the major voice recognition software providers saw that voice applications were finally starting to take off—but that there was a shortage of voice application developers, and no good tools for them to use in authoring those applications. They had their own internal authoring tools, but recognized that they were far too clumsy for the less sophisticated new developers entering the field. They asked Cooper for a high-level vision of how their authoring tools should work.
The solution
I defined a whole new programming environment with a clarified version of the user-facing object model.
I defined a new overall structure for the interface that made it possible to easily manipulate the structure of a voice application in a visual programming environment. This included a novel interface for visualizing (“auralizing”?) the flow of possible dialogues with the application, showing a potential conversation textually, with the ability to change the potential user's input at any juncture. (Clicking on the illustration below will expand out a sample progression showing the dialogue visualizer in action.)
The programming interface I designed showed the potential dialogue linked to the software elements driving it, so that developers could flip from the dialogue to the software structure and back again in order to easily debug the system's behaviors.
What's cool
In just a few weeks, I was able to enter an unfamiliar domain and provide a novel and elegant interaction design solution suitable to the unique challenges of the domain.
