No abstract

We present Wish, a system that uses expert members of an online crowd to amplify a user’s ability to carry out complex creative tasks.  When presented with an incomplete or rough draft of creative content (for example, writing, programming, or art), Wish finds and recruits an expert to suggest a possible realization of the user’s vision. Wish contributes a new approach to crowdsourcing complex work. Rather than combining pools of unskilled workers to producing complex output, Wish uses the crowd to identify and recruit a pool of experts, then assigns a single high-quality expert to carry out the task.  This approach retains the benefits of using a crowd — scalability, speed, correctness, and responsiveness — while simplifying the process of crowdsourcing complex work. We demonstrate Wish in the context of three prototype tools.  Draft enables users to consult a crowd of authors for suggestions on writing.  Hack lets developers convert pseudocode into working code by pulling results on demand from a crowd of programmers. Sketch lets individuals convert rough sketches into fully-refined art by consulting artists.  We illustrate how novice creators can use Wish to amplify their ability, how expert designers can use Wish to explore design spaces and improve the speed of creation, and how new users can use Wish to gain feedback from experts.

We present Euclide, a multimodal system for live animation of a virtual puppet that is composed of a data glove, MIDI music board, keyboard, and mouse. The paper reports on a field study in which Euclide was used in a science museum to animate visitors as they passed by five different stations. Quantitative and qualitative analysis of several hours of videos served investigation of how the various features of the multimodal system were used by different puppeteers in the unfolding of the sessions. We found that the puppetry was truly multimodal, utilizing several input modalities simultaneously; the structure of sessions followed performative strategies; and the engagement of spectators was co-constructed. The puppeteer uses nonverbal resources (effects) and we examined how they are instrumental to talk as nonverbal turns, verbal accompaniment, and virtual gesturing. These findings allow describing digital puppetry as an emerging promising field of application for HCI that acts as a source of insights applicable in a range of multimodal performative interactive systems