In this article, we propose two pointing techniques based on automatic scanning and requiring only a single input switch for the user. The first technique is an improved version of the “Button” technique which was the best of the two versions previously presented in [XX]. The second pointing technique (CS³) is based on circular scanning of the screen, first by an arc, then by a line. First results show that our second technique, CS³, was slower than iButton, but the participants also made fewer pointing errors with this technique. So if we take throughput into account, our two techniques are equivalent in terms of performance.

In this article, we propose WordGlass, a system of additional keys that are dynamically added as the user types. Each key proposes one of the most likely words. The keys have been spaced out on the keyboard so that one additional key can be displayed above and another below the last key pressed, without obscuring the keys already present. First results show that, although typing speed is slightly higher with WordList than with WordGlass (1.3 cps versus 1.25 cps), WordGlass has some benefits: the prediction use rate is much higher with WordGlass than with WordList. The majority of participants said that it was easier to see a word that was close to the pointer than to look beside the keyboard. Moreover, when there were few characters left to type on the current word, participants preferred to continue typing on the keyboard rather than moving to the list. This proximity to the cursor was also reflected in the distance covered by the pointer. On average, participants covered 28.3% less distance with WordGlass than with WordList.

Mouth-operated mouse emulation devices enable an accurate computer control for people with restricted movement of their upper limbs. To be useable even with a limited range of head movement and minimal muscle strength, precise sensor systems are required, which are usually expensive. In this work, we explore the phenomenon of piezoresistivity of SMD thick film resistors and utilize this effect for a low-cost, highly sensitive force sensor for mouth-operated joysticks. A proof-of-concept implementation is presented, including the PCB layout and signal processing strategy. The sensor performance is compared to standard strain gauge measurements, showing minimum detectable forces at the joystick tip of 0.5g (strain gauge) or 10g (piezoresistive measurement). A prototype of a mouth-operated joystick based on the novel sensor boards was evaluated in a user study with 10 participants. In a quantitative evaluation, the efficiency of the pointing device was measured using Fitts' law, showing an average information throughput of 1,51 bits/s, with an error rate of 9,79%. The manufacturing cost of the novel sensor system is only €7.5, as it can be produced almost completely automatically. This drastically reduces the cost of our mouth-operated FlipMouse device (or similar alternative input solutions), making it affordable in low-income settings. All designs have been released under open source licenses.