Dr. Dobb's Journal December 2001
I haven't checked the archives, but it seems like every year about this time, I devote some or all of this column space to nanotechnology. The timing is predictable: It is around this time of year that the Foresight Institute (http://www.foresight.org/) always holds its big Foresight Conference on Molecular Nanotechnology, and by the time I've read the calls for papers and event announcements and postconference reports, nano is very much on my mind. Here, then, is my latest update on all things nano.
First, the history. Or rather, first the gratuitous digression, second the history.
Some time in the last century, two Englishmen wrote a book called 1066 and All That. I'm being appropriately vague about the particulars, in keeping with the philosophy of history espoused by the authors, whoever they were. These alleged historians claimed that their book was the only memorable history of England, because it contained all and only those things that everyone already remembered. History, the authors explained, "is what you can remember." Their book contained just two dates: 1066 and another that I don't remember (but then, I'm not English). In memory (sort of) of whatshisname and whatshisname, here is a history of nanotechnology in three dates.
1959. Physicist, Nobel laureate, wise guy, safecracker, and drummer Richard Feynman gives a humdinger of an after-dinner speech ("There's Plenty of Room at the Bottom," http://www.zyvex.com/nanotech/feynman.html) in which he looks forward to what one might call the ultimate dream of engineering — building things by placing individual atoms or molecules precisely where you want them. In classic Feynman tradition, he wittily describes a thoroughly outrageous notion, then performs a tour de force analysis of its practicality. His conclusion: The laws of nature do not preclude our manipulating matter an atom at a time, and whenever we decide that it is worth the effort, we will actually do it.
1981. Inspired by the published version of Feynman's speech, MIT graduate student Eric Drexler writes a paper called "Protein Design as a Pathway to Molecular Manufacturing" (http://www.imm.org/PNAS.html). This is the first serious presentation of a plan for the creation of devices able to move molecular-scale objects and position them with atomic precision. The paper and subsequent books, notably Engines of Creation, launch Drexler's own career and the careers of a generation of nanoscientists and engineers.
1989. The proof of concept for nanotechnology comes when scientists at IBM's Almaden research laboratories place individual Xenon atoms on a nickel plate to form the IBM logo (http://www.almaden.ibm.com/st/disciplines/nanoscale/).
Today, nanoscience is making great strides, although nanotechnology per se is still in its infancy. There are, though, some practical applications already. (The challenges of nanotechnology are so great that the word "already" should not be taken ironically, even though the field is 20-40 years old.) Some of the practical applications are not being reported as nanotech advances because they fall under a more newsworthy heading — biotech.
Lightyear Technologies, for example, has used nanotech methods to create new chemical catalysts for such purposes as recovering oil from tar sands and purifying water. Nanophase Technologies has what could be (sigh) the killer app: nanotech sunscreen. They've used nanotech methods to improve the properties of zinc oxide used in sunscreen. C Sixty Corp. has a method for using fullerenes, nanotech structures, to kill cancer cells. The field of genomics is focused on using nanoscale probes to illuminate, identify, and mark individual genes in a strand of DNA, and eventually, to move them.
In the area of semiconductor science, developments that reach down to manipulate molecules on the nanolevel are usually reported as if they were just the next inevitable instance of Moore's Law in action. But the world of nanotechnology is the mesoscale, the place where quantum effects compete with macro forces for dominance. Semiconductor technology has been in or near mesoscale for a while now. Ditto for disk-drive surface technology and a number of other computer-related technologies.
Much work in nanotechnology has focused on carbon nanotubes (or buckminsterfullerenes or buckyballs), hollow structures about a nanometer in diameter that can be produced in quantity and used in molecular manufacturing. A Japanese company called "Showa Denka" is currently producing nanotubes at a rate of roughly a kilogram a week. That may not sound like much, but a little nanomaterial can go a long way.
Nano is big. There is a search engine dedicated just to nanotech, a daily news site for nanodevelopments, and nanotech magazines and newsletters. In the last three months of 2001, there were at least 30 significant conferences on nanotechnology and nanoscience, including the IEEE-NANO 2001 conference that I really regret missing, and not just because it was held in Maui.
There are scores of companies doing nanotech work today, hundreds probably, depending on what gets counted. Here are roughly one score of nanocompanies, set up to do everything from pumping out carbon nanotubes to molecular manipulation tools: NanoBio, Nanocor, Nanofactory Instruments, Nanogen, Nanolab, Nanologic, Nanomanufacturing, Nanomat, Nanonics, Nanometrics, Nanophase Technologies, NanoPierce Technologies, NanoPowders Industries, Nanoprobes, NanoSpectra, Nano Storage, Nanostream, Nanostructures, Nanosyn, nanoTITAN, and of course, Nano...and those are just the ones with "nano" in their names.
Much of the groundbreaking nanotech research is being done in university programs. In a quick search, I came up with these programs, but there are more: Arizona State University's NanoStructures Research Group, Brown University's Nano and Micromechanics Laboratory, Caltech's Nanofabrication Group, City University of New York's Nanotechnology and Materials Chemistry, Clemson University's Laboratory for Nanotechnology, Cornell's Nanofabrication Facility, Cranfield University's Nanotechnology Group, Dartmouth College's Molecular Materials Group, Drexel University's Nano Materials Group, Duke University's Thin Films Laboratory, Georgia Institute of Technology's Nanostructure Research Laboratory, UCSB's Institute for Quantum Engineering Science and Technology, Kaunas University of Technology's Research Center for Microsystems and Nanotechnology, Kyushu University's Division of Nanoelectronics, MIT's NanoStructures Laboratory, New Jersey Institute of Technology's Nonlinear Nanostructures Laboratory, Nottingham Trent University's Polymer Engineering Centre, Osaka University's Nanoparticle and Low-Dimensional Compound Group, Penn State's Nanofabrication Facility, Princeton's NanoStructures Laboratory, Purdue's Center for Nanoscale Devices, Queens University at Kingston's nanoPhysics Group, Rice University's Center for Nanoscale Science and Technology, Stanford University's Nanofabrication Facility, USC's Laboratory for Molecular Robotics, University of Texas at Dallas's NanoTech Institute, and the University of Washington's Center for NanoTechnology. Many other schools have nanotech programs under less obvious names, and various research facilities such as Sandia Labs are also big in nano.
Despite some successes, the field of nanotechnology hasn't really answered some of its fundamental questions yet. The vision, articulated by Feynman and Drexler and encouraged by that IBM logo demonstration, is of being able to place molecules, one by one, next to one another, creating new materials and structures with atomic precision. In other words, little machines and levers and gadgets acting as nanoscale analogs of macroscale devices. This probably won't work.
For one thing, you don't get much out of moving molecules one at a time. At that rate, you can't build a structure large enough to be interesting in any reasonable amount of time. Nanomachines would really have to be nanofactories, with millions of nanomachines manipulating molecules one at a time. That raises the question of how you come up with millions of nanomachines and how you assemble them into a nanofactory. Perhaps they are self assembling? That prospect raises all the fears of nanodevices out of control, the nightmare scenarios of science fiction writers. The risks, however, seem at least as real as the possibility of actually creating such self-replicating nanomachines. Or perhaps the metaphor of the machine should not be taken so literally. Maybe instead of literal machines with levers and gears we should think about chemical processes as machines — or biological processes. Maybe nanotech should simply use chemical or cellular mechanisms to produce its better sunscreens and drug-delivery devices.
You can see this uncertainty about methodology in the September 2001 issue of Scientific American, which focuses on nanotechnology. You can also see a sharp difference of opinion between two knowledgeable observers and participants over whether nanomachines are even theoretically possible, and therefore, whether they are worth pursuing. These are not new debates. Drexler laid out most of the questions and alternative approaches years ago in his articles and books. But those are the questions being debated today — which is evidence either of Drexler's vision, the slow progress of research in the field, the difficulty of the problems, or all three. All three, I'd guess. A good source for information and for links to other sources on nanotechnology online is Nanotech Planet, at http://www.nanotechplanet.com/.
Security is on my mind these days, too. Yours, too, probably. I won't remind you why. I just read a good book on security and, while it won't be of use to most DDJ readers, it is so well written and covers the ground so well that I'd like to give it a plug, if for no other reason than to encourage good writing on technical subjects. Maybe you'll recommend it to friends.
Internet Security for Your Macintosh, by Alan B. Oppenheimer and Charles H. Whitaker (Peachpit Press, 2001; ISBN 0-201-74969-6) is primarily for Mac users, and primarily for ordinary users, not system administrators. That shuts out a lot of readers right away. But it necessarily covers a lot of information important to anyone, regardless of platform or level of expertise. The focus lets the authors get very precise about a subject near and dear to their hearts. The authors are in the ISP business and the software business and are known to me personally, so some disclaimers are in order. Open Door Networks, Alan's company, hosts my partner Nancy's web site and online catalog. Alan and Charles discuss in the book, among other things of course, their own software. Oh, and I sometimes play poker with Alan. The disclaimer that actually needs addressing is the one about the authors' discussing their own software. Alan Oppenheimer has some serious credentials in Macintosh networking and security. He once worked at Apple on networking, and networking and security software that he developed has been chosen by both Apple and Symantec for inclusion in their products. The authors could hardly ignore their own software, since it affects Mac security. In fact, Alan and Charles forthrightly point out security problems raised by their own software. Will the book sell a few more copies of Open Door products? Probably. But is its information reliable and balanced? Yes.
The book starts out by talking about physical security, the intelligent management of passwords, and security issues in web surfing and e-mail. It presents a basic overview of Internet protocols and TCP/IP basics, and goes into great detail on how to secure Internet services, including third-party tools (such as Open Door's ShareWay IP tool). It covers viruses adequately and is particularly cogent on the subject of firewalls. The book also includes what is in effect an essay on the dangers of FTP, with suggestions on how to avoid using it and how to minimize your security risk if you do use it.
More advanced topics include a nice discussion of security issues in home networking, work environments, wireless networking (via Airport, of course, but since Airport is a standard, much of this transfers), and MacOS X.
The Airport and OS X information was a little outdated immediately on publication; OS X is a moving target, and revelations about the lack of security of WAP didn't surface in time to get into the book. There is a companion web site with updates, though.
I'm using the book to develop a comprehensive security program for our network here at whatever I call this operation this week. (Nancy's business, Summer Jo's, is half or more of the network; mine, The Prose Garden, is the other half. Just keyboarding that fact suggests to me security issues that I haven't addressed.) I'd recommend the book to any Mac user who has reason to worry about security. That probably excludes anyone using a 56K or slower Internet connection and no LAN; not that such a user has no security concerns, but running a Mac on a slow connection makes one a very unappealing target for mischief makers. The book, by the way, contains no advice on poker.
Computer scientist Alan Turing proposed the classic test for artificial intelligence: If a human being, interacting remotely with an entity in such a way that hides any contextual clues like typing speed or voice inflection, can't tell whether that entity is a human or a machine (or software), the entity is said to have passed the Turing test. If the entity is, in fact, an artificial creation, this is a demonstration of artificial intelligence.
Psychologist B.F. Skinner raised his daughter in an artificial environment, testing on her his theories of learning. Now, neurolinguist Anat Treister-Goren is raising a child in an even more controlled environment, testing her theories of learning on him. "I build his world on a daily basis," she says. Her goal for him is that he should one day pass the Turing test. Because her baby is software. Hal (of course, he would be named Hal) is a computer program that is being raised as a child and taught to speak through experiential learning, the same way human children learn, according to Treister-Goren. Treister-Goren is raising her little Hal at the Israeli company Artificial Intelligence, where she and others are quite serious about the goal — they really hope that Hal will be able to pass the Turing test, maybe in 10 years' time. They plan at least that long a training period for him. And they really are teaching him the way a child learns — Treister-Goren reads to him. Most children don't get their algorithms tweaked by computer scientists between bedtime stories, though.
Of course, it's a little vague just what constitutes passing the Turing test. It shouldn't depend on the gullibility of the human with whom the system is interacting, but how do you control that? In any case, Hal can already fool some of the people some of the time with his 200-word recognition and 50-word spoken vocabulary. That's all acquired; he was born with no vocabulary, just knowledge of the letters of the English alphabet and a preference for rewards over punishments. The ABCs and a value system. He's learning fast.
No word, though, on whether Treister-Goren is teaching Hal to play poker.
DDJ