Humanity: Computing's biggest problem

NEWS Researchers are trying to get a grip on one of the big problems with computers: the human on the other end. A team of computer scientists, academics and others gathered this week at IBM's Almaden Research Centre to swap ideas on how to understand better the ways in which humans interact with machines. Specifically, they met to compare notes on different ways to make sense of how technology is being consumed. Intel, for example, has hired anthropologists to travel the globe, exploring how computing differs in various cultures. Other companies are using new technologies, such as eye-tracking cameras and software that records each click of the mouse to get a picture of what their users are doing. Meanwhile, Web-search king Google finds its own employees are the best guinea pigs on which to test new ideas. "We don't like to do focus groups," said Marissa Mayer, director of consumer Web products for Google. Instead, Google prefers to first ask employees what they think of potential changes and then throw those alterations out to either a test of customers or, in some cases, just try the change on its live Web site. Sometimes, that's how the company realises that it has a bad idea on its hands. That was the case when it offered broadband customers a thumbnail picture of each Web site with its search results. Of those who got the thumbnails, 5 percent of Google users went to the preferences page and turned off the feature. That's a pretty big thumbs-down, considering that typically only 1 percent or 2 percent of people ever view the Google preferences page. The debate over how to improve user interfaces is one of the most pressing in research today. To expand computer use, some researchers are trying to come up with ways to allow people to use their voice, handwriting or other familiar ways to operate computers. And while past gatherings have looked at such radical user interfaces, this year's New Paradigms in User Computing conference was focused on getting a handle on what is understood about the way people use the technology they have. Better tools
Those at the event said they are benefiting from improvements in the tools that track how technology and humans interact. "We're in an age of new microscopes," said IBM's Dan Russell, the conference organiser and senior manager of IBM's User Sciences & Experience Research Lab here. For example, eye-tracking technology, through which cameras capture where someone is looking, has been a popular research technique for decades. But, in the early days of eye-tracking, those being studied were painfully aware they were being tracked because they had to wear a bulky contraption on their head. Now such technology is becoming less obtrusive, with hidden cameras placed facing the user to study eye movements more subtly. At the same time, such tracking can now be combined with software that tracks every click of the mouse, creating a very detailed picture of both what an Internet surfer actually sees as well as what action they take. But despite all the advances in technology, Intel anthropologist Genevieve Bell made the case for leaving the office to do firsthand, in-person research. "You have to go there and hang out," said Bell, who has racked up a fair number of frequent flier miles travelling to homes in India, Malaysia, South Korea, Singapore, Indonesia and China as part of her study of cultural differences within the emerging middle class in Asia. If she hadn't gone to these places, Bell says that she would never have known that in some areas, people are buried with paper versions of their laptops so they will have technology in the next world, while in other spots, people have monks bless their cell phones. Sometimes, Bell said the findings are disturbing, such as the mobile phone faceplates in Malaysia that show a plane crashing into the World Trade Centre. Of more benefit to Intel, Bell has also been finding out where technology resides within people's homes. In China, for example, the television is typically in the living space where someone might entertain a guest, while those who have computers are likely to ensconce them away as far from prying eyes as possible. Such observations might mean China is not a good initial market for a Media Centre PC that combines interactive television with a computer. Others at the conference had other ideas on how to find out what those using technology are thinking. Hewlett-Packard researcher Joshua Tyler presented research showing that a lot can be learned by studying a person's email in-box, even without examining the contents. Tyler and others looked at 900,000 messages from within HP Labs, with the results showing that email discussions closely follow a company's organisational structure or projects, with only limited discussion beyond functional groups. By narrowing in on the patterns within 70 in-boxes, the HP researchers were able to study email threads and how they are distinct from other conversations. Although the size of a conversation on email is often similar to an in-person one, email conversations allow even shy participants to get their say. "Right now, I am talking and it is harder for people to jump in and get involved," Tyler said. But with email, multiple people can respond at the same point in the conversation. "Email is more egalitarian."

---