PonDeFlick: A Japanese Text Entry on Smartwatch Commonalizing Flick Operation with Smartphone Interface

Title of the Paper

PonDeFlick: A Japanese Text Entry on Smartwatch Commonalizing Flick Operation with Smartphone Interface

Paper Information

• Subject Area: Design and evaluation of Japanese text input methods on smart devices
• Keywords: PonDeFlick, Japanese input, software keyboard, Japanese characters, smartwatch

Research Background and Problem

• Identified Problems or Challenges: While QWERTY-based text input methods are standard for Latin-based languages, existing solutions face challenges in handling the input of more complex syllabic characters (kana) and subsequent kana-to-kanji conversion for non-Latin languages like Japanese. Additionally, due to the limited screen size of smartwatches, traditional numeric keyboard downsizing approaches fail to effectively address the issue of insufficient tapping space.
• Significance: The complexity of Japanese input makes it particularly important to design an input method that is efficient, easy to learn, and consistent with interfaces users are already familiar with. Although smartwatches are constrained by their limited interaction capabilities, optimizing the input interface can significantly enhance device usability and convenience in daily life.
• Research Motivation and Related Work:
  • Related studies have proposed various optimization solutions for text input on small-screen devices, such as zooming, reducing the number of keys, and statistical decoding.
  • Existing Japanese smartwatch input methods, such as "BubbleFlick," provide a circular layout but are difficult to learn due to inconsistent gesture directions.
  • This study aims to address the lack of consistency with smartphone input operations in existing interfaces while simplifying the input process.

Solution

• Proposed Method or Solution: PonDeFlick—a Japanese text input method based on a circular layout that enables kana input through touchscreen gestures. PonDeFlick maintains consistency in swipe directions with Japanese smartphone input interfaces and utilizes the entire screen for gesture operations.
• Innovations:
  • Swipe directions are consistent with the numeric keyboard input on smartphones, reducing the learning curve.
  • The circular layout ensures that key areas are large enough to avoid "fat finger" issues.
  • A specialized algorithm for detecting gesture turning points improves the accuracy of input operations for kana.
• Implementation Steps and Key Technologies:
  • Circular Layout Design: Kana and symbol keys are distributed systematically in a circular layout, with a central area for text editing.
  • Gesture Detection Algorithm: By setting minimum distance and angle thresholds for swipes, the system accurately identifies swipe directions and maps them to kana.
  • User Study Design: Performance comparisons between PonDeFlick and other input methods (including existing smartphone numeric keyboards and their modified versions) were conducted to evaluate speed, error rate, and user experience.

Research Results

• Specific Outcomes:
  • In a ten-day user experiment, PonDeFlick achieved an average input speed of 57.7 characters per minute, outperforming the numeric keyboard (KeypadFlick) and its modified version (PonDeSlide).
  • Error rates significantly decreased, reaching approximately 11%—half the initial error rate of the numeric keyboard.
  • The SUS usability score reached 84.4 (rated as "excellent"), surpassing other input methods.
• Advantages:
  • Compared to traditional keyboards and modified input methods, PonDeFlick demonstrated a shorter learning curve and faster speed improvement.
  • The consistency of swipe operations with smartphone interfaces provided significant advantages, especially for users already familiar with smartphone input methods.
• Experiment or Evaluation Results:
  • User data showed higher text input speeds and lower error rates, with significant improvements in operational smoothness and usability.
  • Two key findings: maintaining consistency in swipe directions with smartphone interfaces enhances user learning efficiency, and the circular layout performs well in terms of usability and input accuracy.
• Limitations and Future Directions:
  • The current experiment duration was relatively short, preventing further observation of whether the learning curve fully stabilizes over the long term.
  • PonDeFlick is primarily designed for circular smartwatches, and future work needs to address layout adaptation for square screens.
  • Further exploration is needed to determine whether swipe-based operations can be extended to input interfaces for other non-Latin languages.

Conclusion

PonDeFlick, as a novel Japanese input method, demonstrates exceptional performance and user experience by maintaining consistency with smartphone operations and optimizing gesture input on smartwatch screens. Experimental results show that it surpasses existing solutions in terms of speed, error rate, and user satisfaction. Additionally, the study highlights how integrating interface design with user familiarity can significantly enhance input efficiency. Future research will focus on optimizing its adaptability to different devices and evaluating its long-term performance.