Section: New Results

HCI models, theories, and frameworks

Participants : Géry Casiez, Alix Goguey, Stéphane Huot.

Pointing is one of the most common and frequent action made with any interactive system whether it be a desktop computer, a mobile device or a wall-size display. Although it has been extensively studied in HCI, current pointing techniques provide no adequate way to select very small objects whose movements are fast and unpredictable, and theoretical tools –such as Fitts' law-– do not model unpredictable motion. To inform the design of appropriate selection techniques, we studied human performance (speed and accuracy) when selecting moving objects in a 2D environment with a standard mouse. We characterized selection performance as a function of the predictability of the moving targets, based on three parameters: the speed (S) of the target, the frequency (F) at which the target changes direction, and the amplitude (A) of those direction changes. Our results show that for a given speed, selection is relatively easy when A and F are both low or high, and difficult otherwise [22] .

In spite of previous work showing the importance of understanding users' strategies when performing tasks, HCI researchers comparing interaction techniques remain mainly focused on performance. This can be explained to some extent by the difficulty to characterize users' strategies. To alleviate this problem, we introduced new metrics to quantify if an interaction technique introduces an object or command-oriented strategy, i.e. if users favor completing the actions on an object before moving to the next one, or in contrast, if they are reluctant to switch between commands [21] . Through a study comparing two novel interaction techniques with two from the literature, we showed that our metrics allow to replicate previous findings on users' strategies concerning the latter.

To our knowledge, there are no general design and evaluation methodologies available for the development of digital musical instruments (DMI). One reason is the large diversity of design and evaluation contexts possible in musical interaction, e.g. is this evaluation done from the perspective of the DMI designer/manufacturer, the musician playing it, or the audience watching it be performed? With our collaborators of the MIDWAY associate team, we have analyzed all papers and posters published in the proceedings of the NIME conference from 2012 to 2014 [16] . For each publication that explicitly mentioned the term “evaluation”, we looked for: a) What targets and stakeholders were considered? b) What goals were set? c) What criteria were used? d) What methods were used? e) How long did the evaluation last? Results show different understandings of evaluation, with little consistency regarding the usage of the word. Surprisingly in some cases, not even basic information such as goal, criteria and methods were provided. Beyond the attempt to provide an idea of what “evaluation” means for the NIME community, we pushed the discussion towards how we could make a better use of it and what criteria should be used regarding each goal.