X
Innovation

Rupert Goodwins' Diary

Wednesday 23/8/2006 With the exception of certain Linux fans, smell has yet to feature strongly in the world of IT. The eyes and ears are catered for superbly, with high-resolution moving graphics and synthesised audio getting closer to reality every day.
Written by Rupert Goodwins, Contributor

Wednesday 23/8/2006

With the exception of certain Linux fans, smell has yet to feature strongly in the world of IT. The eyes and ears are catered for superbly, with high-resolution moving graphics and synthesised audio getting closer to reality every day. Touch has its place too, with a steady stream of haptic devices appearing to stimulate and utilise our micron-sensitive fingertips

But smell and its somewhat moister cousin taste? It has a certain crude role to play in diagnosing dead hardware — most engineers are more than familiar with the delicate savour of a scorched resistor, or the pungent pong of an overcooked plastic-packed chip. And anyone who's ever enjoyed the full-on rotten-fish stench of an exploded electrolytic capacitor will know just how powerfully that particular sense can be engaged by electronics.

Yet as an interface, it's been fallow ground. No longer. Down in the depths of the Tokyo Institute of Technology, the Nakamoto Labs have invented a device that can not only produce computer-generated smells but can record and analyse ambient scents as well. It's a full-on odour recorder.

Actually, it's not that full on — it can only work for fruit, and only three fruit at that: orange, apple and banana. But it's the principle of the thing that's exciting: it breaks down mixtures of airborne chemicals by passing them over a nanotech silicon chip with individual molecule receptors, then stores what it finds in a database. Later, that data is used to run a mixing machine that takes nine base compounds and carefully mixes them together to create a simulation of the original smell.

There are major problems before this becomes a general-purpose device. The second biggest is that nobody actually knows how we smell things. We have a few hundred different smell receptors, but how they're triggered by the appropriate molecules — and how that information is then turned into our perception of smells — is still a matter of great debate. It may be possible to synthesise many smells from a small palette of base compounds, or it may need thousands of individual drops of gunk to work.

The biggest, though, is why on earth would you want to? It's a nice thought — briefly — to imagine having your home filled with the scent of a lavender field at dusk, or a tropical forest in bloom. We've all been places where we've been surprised and delighted by smell, as well as had memories forcefully triggered by some half-remembered whiff. But it's not something you need built into your laptop.

I'll only be impressed when we get past the idea of mixing molecules together and get onto proper by-atom synthesis. When we can draw what we want on the screen and have it pop out of the output port seconds later, then we'll be getting somewhere. I'm not sure it'll be used to sythesise Chanel No 5 — but it will create a hell of a stink.

 

Editorial standards