Computerworld

More pleasant surprises, please

Travelling on vacation this week, I was pleasantly surprised a few times by technology. On its first jaunt far from home, my new cellphone made my watch obsolete by sensing time zones and automatically adapting.
  • Jon Udell (Unknown Publication)
  • 07 September, 2003 22:00

Travelling on vacation this week, I was pleasantly surprised a few times by technology. On its first jaunt far from home, my new cellphone made my watch obsolete by sensing time zones and automatically adapting.

One night in an historic inn far from the world of 10-buck-a-night hotel broadband, an 802.11 signal came wafting in on the breeze. Moments such as these are gifts from a future in which technology just seems to do the right thing, quietly and unobtrusively.

But such moments are few and far between. As the novelist William Gibson famously said, "The future is here, it's just not evenly distributed." Too often, our expectations are raised only to be shattered. Why, for example, does the travel search site Expedia ask you to specify the number of children travelling with you, and even their ages, and then proceed to show you hotel rooms that can't accommodate the kids?

When I'm travelling on business, I get to bypass this nonsense. I just call up IDG Travel and talk to Bruce or Michael. These guys can hack through the plane/hotel/rental-car system like hot knives through butter. It's true they have access to privileged information, but I'm sure they could also make better use of public information than you or I. Their ability to recognise and exploit patterns is what makes them so effective.

From science fiction we've absorbed the idea that software systems would behave like Bruce and Michael. In real life that doesn't happen. Arguably it can't, because artificial intelligence is just a sci-fi fantasy. And arguably it shouldn't, because we want human expertise to remain a marketable commodity. But between the dumb software we use and the smart software we imagine lies an important middle ground waiting to be occupied.

Consider email. It gets a bad rap at the moment, deservedly so. Almost all of the email I picked up while travelling was malicious junk generated by the Sobig.F worm. For me this was a non-event, because SpamAssassin was effective in this case. But if I'd had to delete messages by hand, as it seems many folks did, SpamBayes would have quickly recognised the pattern and automated the task.

In the wake of the Sobig.F flood came the predictable reactions: email is polluted, broken, a dead medium. For certain purposes that's true. Writing for a feature on spam I argued that instant messaging is a better solution for real-time communication, as is RSS for bulk notification. But let's not kid ourselves. Factor out spam completely, as my current combination of filters has done for me, and you only do away with a problem. There remains a vast unexploited opportunity.

Thanks to the SpamBayes plug-in, my email client does a great job of learning which messages I do or don't want to see. But it should do so much more. My work life is organised around groups and projects that form and evolve in fluid ways. Email folders and filters help me manage these activities. I want to be pleasantly surprised by software that notices when message patterns indicate the formation of a group or project, and volunteers to set up folders and filters for me. Likewise, I want to be pleasantly surprised by an RSS newsreader that notices how I save and organise items from my subscribed feeds. No breakthrough in artificial intelligence is needed to make this happen. We do the pattern recognition ourselves, quite naturally, as we process our information flows. If software paid more attention to what we attend to, and how, there could be more pleasant surprises.

Udell is lead analyst for the InfoWorld Test Center.