In Search of Text Writing Methods for Off Desktop Computing: ATOMIK and SHARK
		Shumin Zhai – IBM Almaden Research Center


Abstract

The rise of pervasive computing in a variety of forms has set off the 
search of efficient text input methods alternative to the traditional 
typewriter keyboard. This seminar reviews projects my colleagues (B.A. 
Smith, P-O. Kristensson et al) and I have undertaken at the IBM Almaden 
Research Center in this pursuit.  We first designed a virtual keyboard 
for stylus typing. To minimize the stylus movement time, we optimized 
the virtual keyboard with a Metropolis algorithm in which the keyboard 
was viewed as a "molecule" and each key as an "atom". The "atomic" 
interactions among all of the keys drove the movement efficiency - 
defined by all Fitts' law movements between every pair of keys weighted 
by the statistical frequency of the corresponding letters - to the 
minimum.  We also tuned the layout to obtain alphabetical tendency and 
high connectivity of the most common words. The result is the ATOMIK 
Keyboard (Alphabetically Tuned and Optimized Mobile Interface Keyboard).

We are currently exploring a shorthand gesturing system over the ATOMIK 
keyboard.  For each word, the SHARK (shorthand aided rapid keyboarding) 
system defines a shorthand graph on a stylus keyboard - a "sokgraph" - 
based on the movement pattern of the word on the ATOMIK layout and 
recognizes user's gesture accordingly. I will present the key 
rationales of SHARK: (1) the expressive power of the stylus, (2) high 
efficiency stemmed from layout optimization, (3) duality of gesturing 
and stylus tapping, (4) scale and location independent pattern input, 
(5) the power of Zipf’s law, and (6) skill transfer from visual 
feedback based low user performance to recall based open-loop high user 
performance. I will briefly review the SHARK pattern recognition 
methods and an empirical feasibility study of SHARK.

If time permits, I will also present snapshots of other lines of 
research on input and interaction at IBM Research, including our work 
on laws of action  (Accot & Zhai) and eye-gaze sensing based interfaces 
(Zhai, Morimoto& Ihde; Wang, Zhai &Su).


			Shumin Zhai's Bio

Shumin Zhai  is a Research Staff Member at the IBM Almaden Research 
Center. His work at IBM spans three categories:  basic research, 
invention and evaluation of advanced interfaces,  and product 
innovation and development. He has done research and published papers 
on computer input control devices, 3D interfaces, quantification of 
coordination, laws of action, eye-sensing based interaction, text 
input, and the research and development of the IBM Scrollpoint Mouse. 
He holds over a dozen US patents. He is currently on the editorial 
boards of four journals including the ACM Transactions on Computer 
Human Interaction and the Human-Computer Interaction journal . He has 
served on various conference program committees and also served as 
co-chair of the HCI professional interest community (PIC) at IBM 
Research. He has been a visiting professor at the Royal Institute of 
Technology in Stockholm, Linköping University, Sweden; and University 
Paris-Sud, France and has also taught graduate classes at Stanford and 
other universities.  He received his Ph.D. degree from the University 
of Toronto.