In the early 1990s, Japanese economist, Yukio Noguchi, was overwhelmed by the effort of organising the papers that crossed his desk every day, so he did something drastic: he stopped trying. He began throwing every file, regardless of its contents or type, into the same box. Any time he pulled out a file to use it, he didn’t even bother to put it back where he’d got it. He just put it back in the front of the box. And a strange thing happened – his life got easier.
Tidiness is thought of as a virtue and mess as a vice, but the reality is there is a powerful argument in the other direction. In fact, better than an argument, even, a proof. And it comes, perhaps unexpectedly, from computer science, a discipline we think of as one of the tidiest of them all. The algorithms that computers use to manage their time and space turn out to be surprisingly useful in human lives.
Noguchi had stumbled on to the same thing that computer scientists were just discovering themselves. Sometimes mess isn’t just forgivable, it’s optimal. At Bell Laboratories, researchers Daniel Sleator and Robert Tarjan were studying the mathematical properties of “self-organising lists”, a data structure where, like a filing cabinet or a stack of paper, the information is arranged in a single row, from front to back. The closer something is to the front, the faster you can find it.
In what is now the definitive paper on self-organising lists, Sleator and Tarjan looked at a host of different algorithms for how a machine ought to maintain the order of its data for maximal efficiency. They discovered something surprising and wonderful. One algorithm came with guarantees that no other could match (or in their jargon: “No online paging algorithm has better amortised performance”). What was this supreme algorithm? “Move-to-front”– or simply always putting the last thing you touched back at the very front.
The method that Noguchi adopted out of simplicity – or desperation – was, it turned out, the mathematically optimal approach. No strategy (shy of knowing the future ahead of time) can do better. The next time you look at that towering stack of papers on your desk, feel satisfaction rather than shame. By simply tossing everything back on the top of the pile, you’ve been unwittingly following the optimal algorithm all these years.
More broadly than this particular result about piles and files, computer science teaches us to think differently about why we make order in the first place. In computing, the reason to get organised and the reason not to bother can be expressed in the same currency: time. For a computer scientist, sorting is prophylaxis for search. The one and only reason to spend the time required to sort something is that you’ll save at least that much time searching through it later. Many times in human life – just as in computer science – that’s simply not true.
Thinking about sorting as valuable only to support future search would suggest something surprising. Err on the side of mess. Searching something you didn’t sort is merely inefficient. But sorting something you’ll never search is completely wasted effort.
Still, it’s fair to say that in human environments there is more than just efficiency in play. There’s also, for one thing, aesthetics. Redeeming the reputation of mess doesn’t mean we’ll always find it pleasant to look at. Noguchi recalls one colleague taken aback by the condition of his office. “He said it looks very dirty. But it works well. That’s my answer.”