Probabilistic classification

Most of us are overwhelmed by information in our professional lives, and we are hard pressed to organise it in a consistent and coherent manner. This is a profound change from 10 years ago, yet we still cling to a hierarchical classification of data. We desperately roam our computers, trying to organise the overflow of information coming at us from all angles into neat folders. But for most of us it doesn't work. Does the funny e-mail from a client belong in the "customers" folder, or the "fun stuff" folder. Does the powerpoint you made for the board on rising sales belong in the "presentations" folder, the "board of directors" folder, or maybe in the "sales" folder. Or maybe in all three ? The hierarchical way of sorting information is becoming obsolete, it just can't keep up.

It doesn't take a genius to see that there is a major market waiting to be tapped into, and the bets are being placed to replace the old paradigm of hierarchical classification of information. Google Desktop search, A9 and Apples spotlight are all trying to corner this market. All of them by using search algorithms, letting the user search for a word or phrase and returning appropriate results.

But I think that this approach is not good enough to fulfill a users need for information. How do I search for pictures of my motorcycle ? And how can I be sure I have got all of them in my results ? How do I search for funny e-mails ? Can I make a search for tender and loving e-mails from my former girlfriends ? No, not really.

So we need something different, that will present me with the information that I need when I need it. Probably in a graphical manner, since this is the best way of presenting a lot of information in a limited space.

What will it be ? How will it work ? I don't know, but I am thinking hard about it, the stakes are high.