Smart space automation has, for decades, proposed a vision for enhancing the quality of human life. The Internet of Things (IoT) is one of the technologies addressing challenges and automating tasks that were once solely dependent on human intervention. However, what happens if an IoT system goes into a fault state? How can field-service technicians and even moderately tech-literate end-users without prior system knowledge diagnose and fix problems? In this paper, we investigate how different interface paradigms support diagnostic tasks in such scenarios, by presenting an Augmented Reality-based diagnostic tool and comparing it against a traditional 2D diagnostic desktop interface in a controlled user evaluation (n=25). We find that participants using the situated and embodied AR system complete tasks faster, task accuracy remains similar across interfaces and that the AR-based system is associated with higher physical demand. We interpret these findings as reflecting trade-offs between situated and embodied AR interfaces and symbolic 2D interfaces, and discuss implications for the design of diagnostic tools in smart space environments.

Mobile input methods (IMEs) for text entry in interactive applications, are typically evaluated in lab settings, using phrase transcription tasks as the de-facto procedural element in the validation stage of their development. Transcription tasks offer strong internal validity but weak ecological and external validity, due to the composition constraints imposed on participants. Free composition tasks using image stimuli have been proposed as an alternative, but they lack standardized control to regulate elicited text volume and quality, therefore limiting evaluation internal validity. In this paper, we propose a methodology for rigorous, computational sampling of image stimuli from large image datasets, with an aim to control lexical and semantic similarity in user-generated text during IME evaluations. We further deploy diffusion generative AI models to methodically derive stylistic variations from the sampled images, allowing us to examine the effects of original images, abstractive styles and distractor elements on user input. Our findings from an online image description study with crowdsourced participants ( N = 100) and 3,000 captured input samples, demonstrate that stimulus sampling and restyling can be used a methodological control apparatus for evaluating input methods in lab settings, allowing researchers to systematically control participants’ input density and diversity, mitigating the internal validity challenges of image task evaluations. We present evidence-based guidelines for selecting image stimuli and release our datasets, experiment application, analysis code, and participant descriptions.

Typing on mobile soft keyboards is an acquired skill that users learn through experience. Previous research has modelled learning of keyboard layouts via statistical approximation of the probability of correct recall of a target’s (key) location from long and short-term memory, however, there are no insights as to how the mental model of the keyboard may be formed inside the brain. We construct a theory that could explain this process sufficiently, highlighting the importance of mental landmarks for efficient storage and recall of spatial information. The theory is explored and validated through simulation of a cognitive architecture and comparison with extant empirical data, generating insights that could inform the design of new layouts and facilitate the learning process for novice users.

peer reviewed

Figure 1: Generated phrases across the Valence-Arousal space, from left to right 𝑉 + 𝐴 +