SWIFT: A Short Word Solution for Fast Typing
In this paper, we study one specic problem linked to text input techniques based on prediction and deduction lists; namely, short-word problems. Indeed, while prediction is fast and can be easily made effective for long words (e.g. more than 4 characters)
- PDF / 357,839 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 77 Downloads / 166 Views
Abstract. In this paper, we study one specic problem linked to text input techniques based on prediction and deduction lists; namely, shortword problems. Indeed, while prediction is fast and can be easily made effective for long words (e.g. more than 4 characters), short words take longer to be typed with prediction: the time spent browsing a list-based interaction slows the user down. The present study compares two possible approaches where the user selects inside a prediction list of short words versus tactile exploration (native Voive Over for Apple). Results of our comparative study reveal that our technique reduces the overall typing time and the error rate by 38 % compared to tactile exploration. Keywords: Text input · Visual impairment device · Short words · Word prediction
1
·
Touchscreen
·
Mobile
Introduction
Touch screens rapidly and significantly replace physical keyboards on mobile devices. Hence, text entry is now dependent on software (or virtual) keyboards that are widely used by sighted people, but raise accessibility issues for visually impaired users. These users rely on tactile exploration with vocal feedback of the whole screen for entering text, which is time consuming. In order to solve this problem, different prediction-based techniques or deduction system exist. However, if these systems are generally enhance the speed of text-entry for users, they are inappropriate for entering short words. For example, within the DUCK [6] system, the use of the deduction system increases the input speed for words comprised of 5 characters or more. For shorter words, the use of the deduction system is too costly in time compared to the low number of characters to enter. This is even more a problem that some short words are very frequently used. In this article, we therefore focus on entering short words. Our study aims to improve the typing of such words for use with the DUCK system. The proposed interactions have therefore been designed to be easily integrated and in tune with the rest of the interactions already proposed on this system. Both proposals interactions were evaluated and compared with the text input system traditionally used (like VoiceOver entry system). c Springer International Publishing Switzerland 2016 K. Miesenberger et al. (Eds.): ICCHP 2016, Part II, LNCS 9759, pp. 464–471, 2016. DOI: 10.1007/978-3-319-41267-2 65
SWIFT
2 2.1
465
Related Works Adaptation of the Traditional Keyboard
Text to speech (TTS) synthesis is the most used method to assist text entry on a mobile device coupled with a standard keyboard layout (e.g. QWERTY or AZERTY, depending on the culture). The underlying principle is simple: the user moves his finger onto the keyboard or hits a key, and the keys he hits or he moves across are spoken (see e.g. Apple VoiceOver). When he releases his finger, the corresponding character is inputted. Additional gestures can be used as well. One big problem with that comes from the fact that the keys of a virtual keyboard on a mobile device are very small. Thus, the risk of
Data Loading...
