SHARK ²

TLDR

Zhai and Kristensson (2003) introduced a speed‑writing method for pen‑based computing that uses stylus‑keyboard gestures for familiar words and tapping for others. SHARK² eliminates the need to switch between gesture and tap modes, enabling any word in a 10,000–20,000‑word vocabulary to be entered as a shorthand gesture. The system implements a high‑capacity multi‑channel pen‑gesture recognition architecture, algorithms, and interfaces derived from usage characteristics and human performance observations. The new paradigm allows a gradual, seamless shift from visually guided tracing to recall‑based gesturing, and the system’s key components and performance are reported.

Abstract

Zhai and Kristensson (2003) presented a method of speed-writing for pen-based computing which utilizes gesturing on a stylus keyboard for familiar words and tapping for others. In SHARK²:, we eliminated the necessity to alternate between the two modes of writing, allowing any word in a large vocabulary (e.g. 10,000-20,000 words) to be entered as a shorthand gesture. This new paradigm supports a gradual and seamless transition from visually guided tracing to recall-based gesturing. Based on the use characteristics and human performance observations, we designed and implemented the architecture, algorithms and interfaces of a high-capacity multi-channel pen-gesture recognition system. The system's key components and performance are also reported.