Committee Chair

Wu, Dalei

Committee Member

Liang, Yu; Gao, Lan; Wang, Yingfeng; Huang, Chu-Ren

Department

Dept. of Computational Science

College

University of Tennessee at Chattanooga. College of Engineering and Computer Science

Publisher

University of Tennessee at Chattanooga

Place of Publication

Chattanooga (Tenn.)

Abstract

A central task of computer hardware and software is text entry: sending units of human written language to applications such as word processors and spreadsheets. Input units range from individual lexons (Unicode characters) to lexemes (words and compounds). Standard QWERTY keyboards directly access a limited set of lexons and are used to assemble lexemes one lexon at a time. Non-QWERTY lexons and lexemes often require (e.g.) layout switching, dead-key composition, candidate lists, memorized codes, or relatively sophisticated software input methods (IMEs). This dissertation presents µLex, a framework for gesture-efficient IMEs. µLex IMEs are designed to coexist with application and operating system (OS) functionality, support immediate usability, and enable touch-typing over time. OS and application coexistence is largely via IME-only, supplemental gestural semantics associated with keyboard and mouse longpress, presses of keys or buttons for at least a specified duration. Five IME implementation case studies progress from direct lexon selection to branching lexon selection to orthographic lexeme selection to romanized lexeme selection. µTeclas maps QWERTY cues directly to Spanish letters and symbols, supporting keyboard selection with optional display guidance. µLetras adds a mouse-selectable display. µVietChu addresses Vietnamese text entry, where initial cues correspond to multiple targets; display-guided cue sequences resolve the ambiguity. The framework also accommodates input sets such as German, Esperanto, pan-European lexons, and the International Phonetic Alphabet (IPA). µWords and µ词语 (cíyǔ) ‘words’ extend the framework to English and Mandarin lexemes through display-guided selection. Browser prototypes are constructed from frequency lists and romanization data. English spellings and toneless pinyin provide QWERTY-accessible cues; other spelling and romanization systems provide equivalent cues for other languages. Lexeme frequencies guide access, with more frequent lexemes requiring fewer gestures. Across this progression, the dissertation relates cue/target mapping designs to gestural efficiency (GE), measured by gestemes per Unicode character (GPU) and gestemes per lexeme (GPL), with reciprocal Unicode characters per gesteme (UPG) and lexemes per gesteme (LPG). Cumulative corpus coverage (CCC) measures each evaluated inventory’s frequency-weighted coverage within successive gesture budgets. This research strongly indicates that well-engineered cueing systems can enable lexeme IMEs to achieve substantial gestural savings over lexon IMEs.

Acknowledgments

I gratefully acknowledge the indispensable support and direction provided by my advisor, Dr. Dalei Wu. To my dissertation committee members —Lan Gao, Yingfeng Wang, and Yu Liang — thank you for your time, guidance, and contributions to this dissertation and to the broader project it represents. I also thank Chu-Ren Huang for agreeing to join my dissertation committee and take part in the defense. To the Dean of the Graduate School– Ethan Carver– for his encouragement and guidance. Tomy colleaguesand friends throughout the SimCenter, theComputerScienceDepartment, and UTC at large: thank you for the encouragement.

Degree

Ph. D.; A dissertation submitted to the faculty of the University of Tennessee at Chattanooga in partial fulfillment of the requirements of the degree of Doctor of Philosophy.

Date

12-2026

Subject

Computational linguistics; Gesture recognition (Computer science); Human-computer interaction; Input-output analysis; Multilingual computing

Keyword

input methods; text entry; gestural efficiency; multilingual computing; cueing systems; human-computer interaction

Document Type

Doctoral dissertations

DCMI Type

Text

Extent

xxiii, 166 leaves

Language

English

Rights

http://rightsstatements.org/vocab/InC/1.0/

License

http://creativecommons.org/licenses/by-nc-nd/4.0/

Download

Share

COinS