Thursday, March 17, 2005

Book Review: Fab, by Neil Gershenfeld

After reading the first chapter of this book, it was late and time to go to bed. Reluctantly, I put the book away, turned off the light, and spent the next three hours unable to sleep because of the storm of thoughts it had set off in my brain.

Fab is about how to build things. More than that, though: it's about how - sooner than most of us who think of such things expect - anyone will be able to manufacture almost anything in their garage. The centrepiece of Fab is what Gershenfeld calls the 'fab lab', or fabrication laboratory. It's a combination of computer, 3D printers, laser cutters, and assorted other machine tools for working with almost any material, ranging from wood to plastic to metal. Load in the right software, and a fab lab can build, well, pretty well anything you want it to build; even if it can't build it all in one piece, it can build the components, which can then be assembled.

The interesting thing is that a fab lab can - at least theoretically - build another fab lab. This has, to put it mildly, big implications.

The bulk of the book is divided between introductory pieces on the working elements of a fab lab - subtractive and additive manufacturing, sensors, interfaces, engineering software - and case studies of what actual people have done with the prototype fab labs that Gershenfeld's Centre for Bits and Atoms has set up around the world. One such concerned an art student who made herself a 'screaming bag'; another, a Norwegian herdsman who designed and deployed a wireless network for monitoring sheep and reindeer; yet another an Indian village that built measurement devices to test the quality of milk. Note that none of these people were professional engineers, nor did they have any sort of formal technical education. Not only can fab labs manufacture damn near anything, but ordinary people can use them to do so.

What really grabbed me about this book was how soon this technology is likely to be out there. People like K. Eric Drexler have been talking about self-reproducing, molecular manufacturing machines for twenty years now. The technology has become a science fiction staple, as ubiquotous in any serious treatment of the future as computers and space ships. The assumption, however, has always been that we'll need reasonably mature nanotech in order to make these things; Fab shows that, on the contrary, we can make them right now, using existing technology. They won't build spaceships out of pure diamond, but they'll build damn near anything else, including (as noted above) other fabs. And if a fab can reproduce itself, it can also repair itself, and - logically - incrementally upgrade itself.

Eventually, then, fabs will be nanotech-enabled (and able to make those super-cool diamond spaceships); in the meantime, however, we'll have a good twenty or thirty years to learn how to use these things. This is probably a good thing, because the proliferation of fabs - odds are they'll be everywhere within a decade or so - is going to cause one hell of a shock to the economy. The sheer range of what it's possible to make with a fab, and the fact that once you have one all you need are raw materials (dirt cheap) and blueprints (P2P, anyone?), and you can see how large sectors of the manufacturing industry are going to go tits up.

Fab wasn't perfect, the my problems with it were quite minor. Like most engineers, Gershenfeld spends a lot of time explaining the nitty-gritty, and not a lot speculating about the large-scale implications of the technology. Which is fine, really; the point of the book is to show us all what can be done right now, not to engage in hand-waving pronouncements. Still, in the chapter titled 'Future', he spends half of it discussing von Neumann and the machines named for him. The material is relevent, of course, but it really does belong in a different, earlier chapter. The other - minor - problem I had was with some of the case studies, some of which didn't directly concern use of a fab lab, but rather the use of CAD tools in general in contexts they weren't intended for. But then, fab labs are new, and there probably aren't that many examples to draw on as of yet.

In the end, this was an informative and occasionally even entertaining read, an eye-opening guide to a world-changing technology that most of the species hasn't even heard of yet.

As an interesting addendum, I found a post on EurekAlert! today, concerning researchers at the University of Bath who are looking into getting working, refrigerator-sized household fabricators onto the market within four years. It's spreading.


Post a Comment

<< Home