Wired’s magazine’s recent article on Craig Newmark starts with the same stuff I’ve come to expect from hardcore techies. Optimistic to a fault and with their respective employers weathering the current downturn, I.T. professionals don’t seem to be planning for or discussing future technological eventualities. And why should they? Most are problem solvers, and every problem since 1980 has, apparently, succumbed to the relentless march of progress!
Check this out:
“The Internet’s great promise is to make the world’s information universally accessible and useful. So how come when you arrive at the most popular dating site in the US you find a stream of anonymous come-ons intermixed with insults, ads for prostitutes, naked pictures, and obvious scams? In a design straight from the earliest
days of the Web, miscellaneous posts compete for attention on page after page of blue links, undifferentiated by tags or ratings or even usernames. Millions of people apparently believe that love awaits here, but it is well hidden. Is this really the best we can do?”
At first glance this introductory paragraph discussing Craig’s List seems innocent, but the closer you look the more you realize just how laden with assumptions it really is. Now, I’m not saying we can’t make places like Craig’s List better at providing information… but the interesting thing is that the problems explicitly noted aren’t just technical. They’re humanistic. And, no matter how much new technology we invent, no matter how “smart” it gets, people need to understand that nothing exists in a vacuum.
The bottom line? Technology that isn’t built around people’s behavior will always do a poorer job of conveying information than it should. Here’s why.
Information retrieval essentially can be broken down into querries and returns. When I search Craig’s List it offers ways for me to “query” the system. It then “returns” data according to a specific technical workflow. Some approaches are more efficient than others but all return information. The catch is that every single one of these approaches have to translate a query from a person to the system and then return it back again. No problem, right? Wrong.
So long as users (people) play a role in formulating a system query designers have to take unpredictable behavior into account. Most users of technology are complex, rational (while also equally irrational), and uniformly hard to understand. Certainly, there are a lot of things we can know to plan ahead but there’s also a lot we don’t know.
As best I can tell there are two ways to respond. On one hand developers and information professionals can use technology to work around human limitations and automate previously manual processes. On the other, we can accept that users have to be involved to some degree and plan around roadblocks. A huge range of options exist but the ultimate fact remains is that any heavily automated approach to information retrieval relies on the presumption that… we can predict or model human behavior. I’m not going to say we cant, but at the very least a public discussion needs to occur about *how* we can better develop the tools that find our information. Cory Doctorow pointed out in the early 2000′s that for every bit of quality metadata out there there can also be metacrap; useless or misleading information. My grandparents computer habits prove that the things people do with computers defy all expection.
To really transform websites like Craig’s List it’s going to require more than just cool technology.
Or a sudden unexpected breakthough…
visualizations of communication practices – and I realized just how useful that can be. Most procedures at work are pretty difficult to follow, so I think I’m going to do the unthinkable. I’m going to concede that I’ve been wrong.
