The mouse and keyboard, in combination with standard GUI interfaces such as web browsers, spreadsheets and word-processors have become successful and universally accepted in the last decades. However, as we begin to investigate computing off the desktop, such as with mobile or ubiquitous computing, and to broaden the user base of computers to include people with disabilities, these interaction techniques start to face problems.

This talk will highlight some of those problems and discuss solutions that I have developed over the course of the last six years. Some of the problems I have focused on include putting information in the place of need, developing ambient displays that allow virtual information to have a presence in the physical world, developing toolkit support for handling recognition errors and ambiguity in recognition-based interfaces, developing techniques to support people who have trouble selecting mouse targets, and developing new interaction techniques for severely disabled people for handling very low-bandwidth input. This talk will touch on these problem areas in order, focusing in most depth on my thesis work, the problems inherent in handling recognition errors and ambiguity.