[Adamczyk 04]

Adamczyk PD, Bailey BP. If not now, when?: the effects of interruption at different moments within task execution. of the SIGCHI conference on Human. 2004;6(1):271-278. Available at: http://portal.acm.org/citation.cfm?id=985692.985727.

• abstract: measure effects of interrupting a user at different moments within task execution in terms of task performance, emotional state, social attribution; different interruption moments have different impacts on user emotional state and positive social attribution; design of an attention manager;

• intro: poorly time interruptions can adversely affect task performance; authors identify moments for interruption utilising task models based on event perception research;
• results show that predicted points for interruption consistently produced less annoyance, frustration, and time pressure, required less mental effort, deemed by user more respectful of their primary task;

• discussion: predicted worse condition produced the worst results on the same measures, causing users to rate their experience even more poorly than the random condition
• implications for design: designers should consider alternate modalities for interruption; multimodal interruption schemes;
• interruption lag can cue a rehearsal process before onset of an interruption to help in resuming primary task; however, training users in a rehearsal strategy actually decreased task performance [Miller 2007]. (Window of opportunity: Using the interruption lag to manage disruption in complex tasks. HFES '02.)

[Gluck 07]

Gluck J, Bunt A, McGrenere J. Matching attentional draw with utility in interruption. Proceedings of the SIGCHI conference on Human factors in computing systems - CHI '07. 2007:41. Available at: http://portal.acm.org/citation.cfm?doid=1240624.1240631.

• CPSC 544 notes: One aspect that I felt this article was lacking was the emphasis on the need for interruptions and notifications to be consistent across all applications on a system, and within each application a subset of low, medium, and high-priority interruptions. In addition to this, all applications may not be equally important, which will vary from user to user. For instance, User A may rank their email application program as a high-priority application, while their office suite is ranked as medium-priority, and their media player as a low priority application (a medium-priority interruption from the email application (i.e. incoming mail with no high-priority flag set by the sender) would be more important than a medium-priority interruption from the media player (i.e. notification of a song change)). Meanwhile, User B might not care about email at all, and could have an entirely different ranking of applications (and thus a different ranking of interruptions) - (i.e. they want to receive Facebook Wall Post alerts before anything else...). To a certain extent the application GROWL (Mac / Windows) allows users to set these priorities, and make the interruptions consistent in terms of timing and look-and-feel across all applications.

more...