- Observational and Experimental Investigation of Typing Behaviour using Virtual Keyboards on Mobile Devices
- Henze, Niels, Enrico Rukzio, and Susanne Boll. (2012). Observational and Experimental Investigation of Typing Behaviour using Virtual Keyboards on Mobile Devices. Proceedings of the 2012 ACM annual conference on Human Factors in Computing Systems (CHI 2012), 2659-2668.
- Author Biographies:
- Niels Henze is an associate researcher in the Human-Computer Interaction group. He worked and received his PhD in computer science at the University of Oldenburg in Germany. His research interests include interlinking physical objects and digital information and large scale user studies of mobile phones
- Enrico Rukzio is full-time assistant professor at the University of Duisburg-Essen in Germany. His research focuses on mobile human-computer interaction and mobile computing. He began and currently runs the Mobile HCI reserach group. He received his PhD from the University of Munich in computer science.
- Susanne Boll is the professor for Media Informatics and Multimedia Systems in the computer science department at the University of Oldenburg. She is on the executive board of OFFIS-Institute for Information Technology. Her research interests include field of semantic retrieval of digital media, context-aware and location-based mobile systems, and intelligent user interfaces.
The researchs documented keystrokes on the Android keyboard in order to determine trends that would help correlate with a virtual keyboard. The methodology included a typing game. The design of the game collects a large number of keystrokes from people with diverse backgrounds, and was titled "TypeIt". The game consists of three stages with each stage containing four levels, and each level contains multiple keywords. Each level contains white bubbles with the word that needs to be typed. A picture below describes the three stages, stars, water, and fire, more efficiently
In addition, TypeIt was published on the Android market, and data was collected for three months. In total, over 47 million keystrokes were recorded. The researchers provided significant amounts of analysis such as the position of each stroke with a vertical and horizontal alignment. They also noticed that players are faster when typing at the bottom of the screen while strokes are all skewed to the center of the screen as the users glide over the virtual keyboard.
Next, the researchers proceeded to influence the users behavior. They used subtle techniques such as shifting the touch events by a small density per pixel toward the upper part of the screen. This implementation was used to attempt to account for the skew discovered when users were typing. The results were then split into speed, performance, error rate, and learn-ability to the shifted dots.
Related work not referenced in the paper:
1) "Keyboards without Keyboards: A Survey of Virtual Keyboards" by Kölsch and Turk
2) "Performance Optimization of Virtual Keyboards" by Zhai, Hunter, and Smith
3) "Movement Model, Hits Distribution and Learning in Virtual Keyboarding" by Zhai, Sue, and Accot
4) "Performance optimizations of virtual keyboards for stroke-based text entry on a touch-based tabletop" by Rick
5) "Reconfigurable Virtual Keyboard" by Kanade, Kharat, Raundale, Sangve, and Mane
6) "A Field Comparison of Techniques for Location Selection on a Mobile Device" by Luimula, Sääskilahti, Partala, and Saukko
7) "A mobile-based knowledge management system for “Ifa”: An African traditional oracle" by Folorunso, Akinwale, Vincent, and Olabenjo
8) "User-Interface-Technologies and -Techniques" by Steinhage, Rantzer, Niman, and Dainesi
9) "Technologies for Virtual Reality/Tele-Immersion Applications: Issues of Research in Image Display and Global Networking" by DeFanti, Sandin, and Brown
10) "Spy-resistant keyboard: more secure password entry on public touch screen displays" by Tan, Keyani, and Czerwinski
The related works described above offer some powerful potential emerging technologies that are currently considered for development. Nearly every single research report focused on a novel and intriguing idea that is related in someway to a virtual keyboard. However, one of the distraught associations with the work is that not entirely all selections were unanimously relevant to the research paper of study in this blog. In that manner, the research reports noted above were either parallel in every single methodology with this paper reading, or were a somewhat of a stretch to relate with. Nevertheless, there existed some correlation and an abundant amount of interest. In essence, the main difference that this research report concentrates on is the improvement of speedup and accuracy in typing whereas the other reports stray from this domain.
Evaluation:
To begin, the researchers measured the quantitative effects of their shift with accumulation of speed data. This was all quantitative as a finite time was assigned to each word and the duration it took to type. Next, performance was investigate based on keystroke per second, thus giving an accuracy to each word typed. Aso, error rate was quantified with number of mistypes and adaptability to the dots was measured by the difference with and without the dots. In effective, the researchers set up a completely unbiased methodology to evaluate their results in the most quantifiable terms possible
In essence, this report is the most solid terms of evidence. All data was analyzed without the use of biased behavior. However, once the results were all collected, the researcher did have to use subjectivity to evaluate their numbers. For example, the data returned results with change in speed, performance, and error rate compared to no shift change. Usually, a shift change resulted with each vector gaining or losing a certain percentage. Thus, the researchers used the most positive shift change with dots over the keyboard to advocate a potential shift change in development
Discussion:
The paper presents an intriguing and potentially useful idea. Although it is not exciting, the research presented here still pose a significant efficiency advantage in smartphone applications. Thus, it cannot be ignored. I would consider it novel, but in an atypical way. One must realize that saving a mere few hundredths of a second per word can correlate to uncountable man hours saved over the course of a smartphone's lifetime. The only addition I would have enjoyed seeing in the research would be more drastic keyboard changes. Rather than shifting the density of each keyword, I would have enjoyed see them test keyboard configurations or button interchanges.
3) "Movement Model, Hits Distribution and Learning in Virtual Keyboarding" by Zhai, Sue, and Accot
4) "Performance optimizations of virtual keyboards for stroke-based text entry on a touch-based tabletop" by Rick
5) "Reconfigurable Virtual Keyboard" by Kanade, Kharat, Raundale, Sangve, and Mane
6) "A Field Comparison of Techniques for Location Selection on a Mobile Device" by Luimula, Sääskilahti, Partala, and Saukko
7) "A mobile-based knowledge management system for “Ifa”: An African traditional oracle" by Folorunso, Akinwale, Vincent, and Olabenjo
8) "User-Interface-Technologies and -Techniques" by Steinhage, Rantzer, Niman, and Dainesi
9) "Technologies for Virtual Reality/Tele-Immersion Applications: Issues of Research in Image Display and Global Networking" by DeFanti, Sandin, and Brown
10) "Spy-resistant keyboard: more secure password entry on public touch screen displays" by Tan, Keyani, and Czerwinski
The related works described above offer some powerful potential emerging technologies that are currently considered for development. Nearly every single research report focused on a novel and intriguing idea that is related in someway to a virtual keyboard. However, one of the distraught associations with the work is that not entirely all selections were unanimously relevant to the research paper of study in this blog. In that manner, the research reports noted above were either parallel in every single methodology with this paper reading, or were a somewhat of a stretch to relate with. Nevertheless, there existed some correlation and an abundant amount of interest. In essence, the main difference that this research report concentrates on is the improvement of speedup and accuracy in typing whereas the other reports stray from this domain.
Evaluation:
To begin, the researchers measured the quantitative effects of their shift with accumulation of speed data. This was all quantitative as a finite time was assigned to each word and the duration it took to type. Next, performance was investigate based on keystroke per second, thus giving an accuracy to each word typed. Aso, error rate was quantified with number of mistypes and adaptability to the dots was measured by the difference with and without the dots. In effective, the researchers set up a completely unbiased methodology to evaluate their results in the most quantifiable terms possible
In essence, this report is the most solid terms of evidence. All data was analyzed without the use of biased behavior. However, once the results were all collected, the researcher did have to use subjectivity to evaluate their numbers. For example, the data returned results with change in speed, performance, and error rate compared to no shift change. Usually, a shift change resulted with each vector gaining or losing a certain percentage. Thus, the researchers used the most positive shift change with dots over the keyboard to advocate a potential shift change in development
Discussion:
The paper presents an intriguing and potentially useful idea. Although it is not exciting, the research presented here still pose a significant efficiency advantage in smartphone applications. Thus, it cannot be ignored. I would consider it novel, but in an atypical way. One must realize that saving a mere few hundredths of a second per word can correlate to uncountable man hours saved over the course of a smartphone's lifetime. The only addition I would have enjoyed seeing in the research would be more drastic keyboard changes. Rather than shifting the density of each keyword, I would have enjoyed see them test keyboard configurations or button interchanges.
No comments:
Post a Comment