An Automated Evaluation and Critiquing Aid for User Interface Design

We will leverage our prior work on Adaptive Information Management (AIM) to provide a core reasoning capability usable in a diverse set of tools to aid and assist in the review, design, evaluation and management of user interfaces (UIs). We have a task-linked information representation and associated algorithms for reasoning about the information needs of a set of user tasks, the information presentation capabilities of a set of devices and formats, and the degree of match between the two. These have served as the basis of multiple, successful AIM systems in the past, but within highly restricted domains. The proposed project will explore methods and required modifications for moving our core representation in innovative direction-?making it more interactive and subject to user guidance, applicable to a set of contexts of use rather than a current usage "slice", and capable of providing aiding across a broader set of domains. Substantial modifications to our core task representation and several new critiquing modules will likely be needed. Their design will open the doors for a variety of types of Multi-modal Aids for Interface Design (MAIDs). In particular, the proposed project will focus on designing a tool to evaluate human-designed procedure execution displays against users' tasks needs, and provide design feedback and advice.