Human interruptibilityAs designers of computer systems, we want our products to be helpful, efficient, and customized to our users' needs. So, we build in functionality for dialog boxes, audio status cues, and instant communication between remote users. But we run the risk of making these machines yipping puppies, constantly querying users whether they want to update to the latest software version, check their new mail, or scan for viruses right now. The critical factor is no longer technological capacity, but rather human ability to pay attention to a task amidst the myriad interruptions we get.
Which is why I'm especially impressed with James Fogarty, Scott Hudson, and Jennifer Lai's research on human interruptibility (here's their most recent CHI paper). Their work looks for ways to predict whether "now is a good time" for an interruption based on a handful of factors, such as phone use, whether the door is cracked, or the presence of another person in the room. Their model is more powerful than the average human guesser, even with the most intrusive sensors (camera and microphone) removed. And it appears to translate well across different kinds of workers, such as highly social managers, focused programmers, and interns in shared offices. I'm curious to see how well it could be applied to workers in different countries, or to people who don't use computers as their primary tool, such as traditional artists. Given the model's already-proven flexibility, I wouldn't be surprised if it continued to excel, as long as it were fed a little extra training data for these diverse groups.
Additionally, the paper wisely indicates that people's opinions of interruptions are relative, both for the interrupter and -tee. An automatic sensor system shouldn't unconditionally prevent all incoming alerts at "bad times;" instead, it should allow the requester to choose whether to leave a message or to continue through. Then you get into the infinitely more difficult research area of ego and politeness, which we'll leave to the sociologists.