Assessing Feedback for Indirect Shared Interaction
1 Assessing Feedback for Indirect Shared Interac5on with Public Displays Jorge C. S. Cardoso – email@example.com Rui José – firstname.lastname@example.org
2 Our work • Study and develop an interacAon abstracAon – We are starAng with the concept of widget and adapAng it to the public display seEng • In this paper, we focused on the feedback to user input
Target public display applicaAon 3 environment • Environment for public display applicaAons – MulAple displays – MulAple applicaAons scheduled for each display – MulAple interacAon mechanisms – MulAple interacAon points for an applicaAon • Personal mobile devices, dedicated input devices – Ubiquity of interacAon and informaAon • Not focused only on the public display: • interacAon can happen even if the applicaAon is not currently being shown on the public display • ApplicaAons can have mulAple informaAon outlets: web, mobile device, etc.
4 MoAvaAon • InteracAon is import for the success of public displays, but • It is sAll hard to incorporate interacAvity into public display applicaAons, because • “At present, there are no accepted standards, paradigms, or design principles for remote interacAon with large, pervasive displays.” 
5 MoAvaAon • There are no interacAon abstracAons for public displays that take into account: – the heterogeneous input mechanisms – the mulA-‐user, shared and possibly remote interacAon environment
6 Widgets • Our iniAal set of widgets maintain the graphical aﬀordances of the desktop widgets • But they can be interacted with in a very diﬀerent way – They have explicit, human-‐readable, ids that can be used with several input mechanisms
7 Study • Widgets should provide system-‐level input feedback – But given the mulA-‐user, shared, remote interacAon seEng, we must develop a feedback mechanism speciﬁc for public displays • Independent of the input mechanism • FuncAoning as a sharing mechanism to enAce further interacAon  • The purpose of this study was to assess how well users understood the feedback mechanism
9 Feedback Mechanism • The feedback is shown directly on the public display (inside or near the graphical representaAon of the widget) – It includes a masked idenAﬁcaAon of the user that interacted – It also includes an indicaAon of whether the input was accepted (widgets may be temporarily disabled and not accept the input) • We tested three variaAons – Internal: Uses the text components of the widget – External: Uses a pop up for every input – External-‐CumulaAve: Uses the same pop up for the all inputs within a small Ame period
11 Procedure • We asked parAcipants to interact with mockup applicaAons and interpret the feedback • Using SMS messages to interact – The feedback used the phone number to idenAfy the user • Three groups of users were subjected to the three variaAons • Three parts – Single user interacAng with the display – MulAple users interacAng with the display – QuesAonnaire to assess subjecAve preferences • We asked users to interact and interpret the feedback – And measured the errors in their answers • We also asked them to compare the three variaAons and tell us which one they preferred
12 Part 1 and 2: Error results Part 1 − Single−user Part 2 − Multi−user3.0 8 62.0 41.0 20.0 0 INT EXT INT EXT E−C
13 Part 3 100 INTERNAL EXTERNAL 80 EXTERNAL_CUMULATIVE 60% 40 20 0 BEST WORST
14 Conclusions • The External-‐CumulaAve was clearly preferred and had the best performance • This study was a controlled experiment so we sAll need a real seEng evaluaAon, but the results seem to indicate that, generally, the feedback mechanism is understandable.
15 References •  Bellucci, A., Malizia, A., Diaz, P., & Aedo, I. (2010). Human-‐Display InteracAon Technology: Emerging Remote Interfaces for Pervasive Display Environments. IEEE Pervasive CompuAng, 9(2), 72-‐76. IEEE. doi:10.1109/MPRV.2010.30 •  Brignull, H., & Rogers, Y. (2003). EnAcing People to Interact with Large Public Displays in Public Spaces. In M. Rauterberg, M. Menozzi, & J. Wesson (Eds.), INTERACT’03 (pp. 17-‐24). IOS Press. Retrieved from hpp://dblp.uni-‐trier.de/db/ conf/interact/interact2003.html#BrignullR03