Handheld devices present unique obstacles to user interface design. The standard approach is to simulate a PC environment and apply it on a smaller scale. There are two major problems with this approach. First, the usage model for handheld devices is fundamentally different than that of PC's and laptops. Handheld devices are accessed frequently for short bursts (e.g., checking one's schedule or looking up a phone number), whereas PC's are accessed infrequently for long sessions (e.g., creating spreadsheets or writing memos). This results in different tradeoffs and design decisions.

For example, it becomes critical to minimize the number of steps to perform common functions. The second problem is that PC designs are designed for 640 x 480 resolution. On smaller screens, every pixel matters and control objects designed for larger screens take up valuable real estate that could be put to better use displaying information.

We have designed an operating system specifically for handheld computers that addresses these issues. The presentation focuses less on theory and more on the real-life application of our design, covering the following topics

• Lessons learned from early PDA software design
• PalmPilot design goals and informal metrics used to measure our success.
  Special emphasis will be placed on addressing the conflicting goals of minimizing the number of objects displayed for simplicity vs. maximizing the number of objects for fast access.
• PalmPilot design process