Mjolnir: talks

It has become second nature to "Google" any problem we have with the technology we own. As designers of technology, we can exploit this behavior to learn what people want and need from their products, by virtue of what they search for. In this talk, I will present a technique that transforms web search query logs into forms that enable one to identify usability issues in publicly available products, such as the iPhone, Kindle, and Firefox. Since raw query logs are not made available to the public, I will also show how search query corpora can be approximated using publicly available resources. Finally, I will demonstrate how we can automatically map a search query to the specific features in the interface (commands, controls) most related to that query. This capability enables new interaction techniques, such as dynamic tooltips that describe how a command is actually used in practice, or services that can automatically map equivalent commands between two different applications.

Michael Terry is an associate professor in the Cheriton School of Computer Science at the University of Waterloo, where he co-directs the HCI Lab. His research examines how web-centric data sets (search query logs, actual web pages) can be used to model how people think about and use interactive systems. His work also explores the notion of crowdsourcing interface customization, or user interfaces that allow a user community to easily create and share interface customizations.