Text entry for extended reality (XR) is far from perfect, and a variety of text entry techniques (TETs) have been proposed to fit various contexts of use. However, comparing between TETs remains challenging due to the lack of a consolidated collection of techniques, and limited understanding of how interaction attributes of a technique (e.g., presence of visual feedback) impact user performance. To address these gaps, this paper examines the current landscape of XR TETs by creating a database of 176 different techniques. We analyze this database to highlight trends in the design of these techniques, the metrics used to evaluate them, and how various interaction attributes impact these metrics. We discuss implications for future techniques and present TEXT: Text Entry for XR Trove, an interactive online tool to navigate our database.

Any investigation of learning unfamiliar text entry systems is affected by the need to train participants on multiple new components simultaneously, such as novel interaction techniques and layouts. The Guided Evaluation Method (GEM) addresses this challenge by bypassing the need to learn layout-specific skills for text entry. However, a gap remains as the GEM's performance has not been assessed in situations where users are unfamiliar with the interaction technique involved, here eye-gaze-based dwell. To address this, we trained participants on only the eye-gaze-based interaction technique over eight days with QWERTY and then evaluated their performance on the OPTI layout with the GEM. Results showed that the unfamiliar OPTI layout outperformed QWERTY, with QWERTY's speed aligning with previous findings, suggesting that interaction technique expertise significantly impacts performance outcomes. Importantly, we also identified that for scenarios where the familiarity with the involved interaction technique(s) is the same, the GEM analyzes the performance of keyboard layouts effectively and quickly identifies the best option.

State-of-the-art Virtual Reality (VR) and Augmented Reality (AR) headsets rely on singlefocal stereo displays. For objects away from the focal plane, such displays create a vergence-accommodation conflict (VAC), potentially degrading user interaction performance. In this paper, we study how the VAC affects pointing at targets within arm's reach with virtual hand and raycasting interaction in current stereo display systems. We use a previously proposed experimental methodology that extends the ISO 9241–411:2015 multi-directional selection task to enable fair comparisons between selecting targets in different display conditions. We conducted a user study with eighteen participants and the results indicate that participants were faster and had higher throughput in the constant VAC condition with the virtual hand. We hope that our results enable designers to choose more efficient interaction methods in virtual environments.

Previous work has shown that the mismatch between disparity and optical focus cues, i.e., the vergence and accommodation conflict (VAC), affects virtual hand selection in immersive systems. To investigate if the VAC also affects distal pointing with ray casting, we ran a user study with an ISO 9241:411 multidirectional selection task where participants selected 3D targets with three different VAC conditions, no VAC, i.e., targets placed roughly at 75 cm, which matches the focal plane of the VR headset, constant VAC, i.e., at 400 cm from the user, and varying VAC, where the depth distance of targets changed between 75 cm and 400 cm. According to our results, the varying VAC condition requires the most time and decreases the throughput performance of the participants. It also takes longer for users to select targets in the constant VAC condition than without the VAC. Our results show that in distal pointing placing objects at different depth planes has detrimental effect on the user performance.

In eye-gaze-based selection, dwell suffers from several issues, e.g., the Midas Touch problem. Here we investigate saccade-based selection techniques as an alternative to dwell. First, we designed a novel user interface (UI) for Actigaze and used it with (goal-crossing) saccades for confirming the selection of small targets (i.e., < 1.5-2°). We compared it with three other variants of Actigaze (with button press, dwell, and target reverse crossing) and two variants of target magnification (with button press and dwell). Magnification-dwell exhibited the most promising performance. For Actigaze, goal-crossing was the fastest option but suffered the most errors. We then evaluated goal-crossing as a primary selection technique for normal-sized targets (≥ 2°) and implemented a novel UI for such interaction. Results revealed that dwell achieved the best performance. Yet, we identified goal-crossing as a good compromise between dwell and button press. Our findings thus identify novel options for gaze-only interaction.