Gustav Mahler's Symphony No. 4 in G major

http://en.wikipedia.org/wiki/Symphony_No._4_(Mahler)
performed by the Vienna Philharmonic Orchestra (Wiener Philharmoniker), conducted by Leonard Bernstein

• 1st movement (1, 2)
  
• 2nd movement (1)
  
• 3rd movement (1, 2, 3)
  
• 4th movement (1)

Cell-level simulation and hobbyist participation

These are a few simulators for biological cells and processes.

• http://www.e-cell.org/ecell/
• http://www.vcell.org/
• http://synbioss.sourceforge.net/

A lot of the important things that happen in medicine are happening at the cellular level. Cell-level simulators might provide a way for large numbers of hobbyist medical researchers to construct and test hypotheses. The most promising hypotheses might be testable in real biology laboratories, and the results could be fed back to improve the accuracy of the simulators.

I'm not sure this would be an effective strategy for hastening the pace of medical progress. My intuition is biased because I've spent the last fifteen years working with open-source software (Linux, Apache, etc). I recognize that competition and profit are also powerful forces driving the rate of innovation, and that these seem to work best when people aren't sharing information so readily. The software/internet world has seen lots of progress in the last ten or twenty years, and it seems that a mixed environment with both open-source and closed-source approaches has pushed things along well.

Software is difficult but biology is much more difficult. At least it looks that way from my software engineer's point of view. The depth of expertise required for meaningful contribution to medical knowledge will likely exclude most would-be contributors. I don't know what to do about that. Perhaps cell simulators and on-line information can make that expertise more accessible.

Participation by hobbyists has become a very big part of the astronomy community. Maybe there is a legitimate place for hobbyists in the field of medical research.

First post

Here is today's nifty piece of medical progress, an advance in the fight against cancer. A couple years ago, an unfortunate woman died of acute myelogenous leukemia, leaving behind samples of her cells, some healthy and some cancerous. A team at Washington University in St. Louis was able to sequence the DNA from the cells and compare her healthy DNA to her cancerous DNA. This became possible because the price of DNA sequencing equipment has come down by a very large factor in recent years.

They identified ten point mutations that differentiated the sick cells from the healthy ones. Two of the mutations were already known from earlier research, the other eight mutations were previously unknown. The team is continuing to study differences in the non-coding DNA as well, and they are also preparing to apply the same sequencing methodology to other cancers.

Because they were using DNA samples all from the same person, there would be very few differences among the healthy cells, just the infrequent cell-to-cell mutations that might occur in an average healthy person's body. So they had a good solid statistical baseline that made the ten cancer-related mutations really clear.

It may be years before this translates into clinical practice that saves lives. But it's nevertheless an important advance. It's something that has never been done before, and it does bring to light a few new facts. It looks strongly like point mutations are the cause of at least some, possibly all, cancers. We strongly suspected that before but this almost proves it. One of the ten mutations was present in only a fraction of the cancerous cells, suggesting that the mutations typically occur in a particular sequence, with the last one finally making the cell dangerous.

I'm interested in what social and economic factors could most hasten the rate of medical progress. My reason for this interest is simple: I'm not young any more. I'm curious about whether the development model that has been so successful for open-source software could somehow be applied to quicken the pace of medical progress.

Beethoven on Youtube

I've long been a fan of Beethoven's Violin Concerto (1, 2, 3). Here it is played by Yehudi Menuhin.

I also like some of the symphonies. Here's Herbert von Karajan conducting the Seventh Symphony, which I think doesn't get enough attention by comparison to the Third (1, 2, 3? 4?) or the Fifth or the Ninth (1, 2, 3a, 3b, 4). Why is it the only even-numbered Beethoven symphony you ever hear is the Sixth? In fact, none of Beethoven's power-of-two symphonies (numbers 1, 2, 4 and 8) get much airplay. Wierd.

Bach on YouTube

Brandenburg Concerto number 3, probably my favorite: first movement, second movement, third movement. For the third Brandenburg, Bach didn't really write a second movement. He just wrote a couple of chords and allowed the musicians to improvise whatever they wanted within that minimal harmonic constraint. Different groups do different things with that freedom. The first time I heard this concerto was Walter Carlos's rendition on the Switched-on Bach album back in the seventies, which included a lot of interesting sounds that people now associate with old bad sci-fi movies. But at the time, Carlos was one of the first explorers of electronic music and there wasn't yet an esthetic for it. In a later recording Carlos did something a bit more conventional, a minimal expansion on Bach's two chords with just a few flourishes.

One thing I never quite got about the first Brandenburg (first, second, third, fourth movements) is some funny work in the horns in the first movement. There are points where they just seem off-tempo with everybody else. When I first heard this I assumed the musicians had gotten lost. But now I'm hearing it in this second recording, so I have to conclude that Bach wrote it that way. Maybe he was trying to make sure the listener was awake? Perplexing.

I once read a review of the sixth Brandenburg (first, second, third movements) suggesting that it was a musical description of goings-on in the Bach household. Bach had lots of kids, all presumably running and bouncing about as kids will do, and this is a very busy concerto with a lot happening. So that might be what Bach had in mind, and it especially sounds that way in the third movement which has a real bounce to it. In this recording the cellos (maybe basses? I'm never sure) at the right end seem to have many more than four strings.

Multimachine

Multimachine, built by Pat Delany of Palestine, Texas, is an inspiring project. It is...

a humanitarian, open source machine tool project for developing countries... The MultiMachine all-purpose machine tool can be built by a semi-skilled mechanic with just common hand tools. For machine construction, electricity can be replaced with "elbow grease" and the necessary material can come from discarded vehicle parts. What can the MultiMachine be used for in developing countries?
AGRICULTURE...
WATER SUPPLIES...
FOOD SUPPLIES: Building steel-rolling-and-bending machines for making fuel efficient cook stoves and other cooking equipment...
TRANSPORTATION...
EDUCATION...
JOB CREATION...

The project is open source and thoroughly documented. It uses commonly available pieces. It seeks explicitly to address the needs of the developing world. It recognizes the work people did in this area (1, 2) in years past. Cool stuff. We have all kinds of Industrial Revolution era mill buildings in the greater Boston area and this would fit right in.

Frostbot, the work of Brian Schmalz, is another food fabber. It's designed to frost cookies. The CNC mechanism is from Fireball CNC. Brian's other tinkerings include a cool USB bit-whacking board available at Sparkfun.

3d printer project at Victoria University of Wellington School of Design

There was recently a design contest at VUW School of Design to create inexpensive 3d printers. Apparently Ponoko had some involvment, possibly a sponsorship. I found the prettiest printer to be the Equinox, which also was designed to be environmentally friendly, using a lens to focus sunlight to dry recycled paint as a printing process. I was going to say the printer itself looks like an astrolabe, but it really looks like an armillary sphere, a sort of 3D astrolabe.

There's a high-res photo (low-res version below) that shows some of the mechanics, which were laser-cut on Ponoko. I am very much hoping that the university and/or students will publish the plans for the Equinox. It's very cool that people can use a service like Ponoko to build their own printers.

Very slow progress on my CNC mill. I finally purchased the 3-axis Xylotex controller. It is my hope to connect it this evening and conceivably mill a piece of wood. So I need to move a PC to where the mill is, run a network cable, load the PC with EMC and configure it, and set up the shop-vac to collect sawdust. I'll mount the Xylotex board and power supply and fan on the side of the mill, but that's for later.

Penny wise, pound foolish

I got the idea that I should try to design and build my own electronics. I've done electronics design before, including microcontrollers and FPGAs and the like, but I have little experience with power electronics. That, and I'm impatient. The upshot is that after wasting about three weeks and a few hundred dollars in trying to control stepper motors, I'm not much further ahead. Here is the affordable pre-packaged solution (which had been recommended by the guy who sold me the mechanics) which I should have used from the start:

• Xylotex three-axis controller with power supply
  $205 plus shipping
• Xylotex four-axis controller with power supply
  $235 plus shipping
• Xylotex FAQ

So I'll plan to pick up a 3-axis controller and run it with LinuxCNC. Apparently TurboCNC is also very popular but I'm not about to run DOS on any machine that could be running Linux.

The mechanics cost about $300 including shipping. The steppers cost $75 (I got them from RRRF). This stepper controller will run maybe $225 with shipping, so the whole thing is $600. That's reasonable. Obviously it doesn't include waste.

I'm thinking it would be fun to fool with Python code that generates G code and sends it to the CNC. I could develop a repertoire of programmatically defined shapes.

RepRap replicates, and Will gets a New Toy

On the left is Adrian Bowyer, the University of Bath professor who started the RepRap project. On the right is Vik Olliver, the most active RepRap builder on the planet. The two machines marked "parent" and "child" are RepRap 3D printers with the interesting relationship that the "child" was mostly built by the "parent". This is a HUGE STEP toward Bowyer's vision wherein RepRaps make more RepRaps and humans benefit. This will do for physical goods what the GPL and Linux and Apache have done for software.

My own news is, at least locally, equally exciting. My CNC mill has finally arrived! And I also got an Arduino controller. I've got my stepper motors from RRRF, and a Harbor Freight router is on the way. It's going to take time to put everything together, and of course there's very little spare time in the life of a modern adult.

Once the CNC mill is up and running, I plan to work on a scheme for swapping out the router and swapping in an extruder for thermoplastic. By that time the RepRap guys will be doing even better than they're doing today, so I will benefit from their stuff. Maybe I'll end up making an actual RepRap before I'm through.

RepRap: Big step up in print quality!

This posting on the RepRap blog shows the massive progress these guys have made recently in their printing quality. The progression is clearly visible in this photo of some door handles. The most recent incarnation is the work of a guy known as "Nophead", with his own blog describing his work. His machine uses a RepRap extruder on a purchased CNC table rather than the RepRap 3D platform, which made me think that the RepRap platform must be the reason for the less-than-commercial-grade print quality. I asked him about this in a comment, and he replied that the improvements were:

• his extruder has a shaft encoder to control the speed precisely
• he has temperature control to +/- 3C
• he doesn't have any comms delays (I don't know the architecture well enough to know exactly what he means here)
  
• he runs his head faster so as to stretch the filament down to 0.5mm.
• careful choice of printing material
  

To conclude, he says "All these things can be sorted out on Darwin [the current RepRap prototype] so I expect its prints to be this good in a month or two." That's a very cool thing. It's wonderful to see such progress.

Within just a year or two, RepRap will be much further along in terms of both quality and ease of use, and it will be affordable for small clubs in high schools and colleges all over the world, and large numbers of individual hobbyists. By then it will probably print multiple materials including conductive ones, so you'll be able to embed circuitry in a widget. Today one of the big killer apps for 3D printers is little action figures based on avatars from Second Life and similar games, but when 3D printers really are ubiquitous, people will move on to far more interesting apps that I can hardly imagine.

Let me not forget this very nice list of a lot of different commercial and hobbyist 3D printers.

Still waiting for my CNC mill platform, the eBay fellow has been getting a huge volume of business and his shop is a bit swamped. I've been getting a bit more organized with the electronics, including resuscitating an old FX2 board design, and I've ordered some stepper motor driver parts that should arrive soon.

Affordable CNC gadgets

CNC has existed as a hobbyist pursuit much longer than 3D printers have been. I finally broke down and purchased one of these on eBay. It will take a couple weeks to arrive, and the one I got did not include stepper motors, couplers, or motor drive electronics. Those are things I'd enjoy doing myself anyway, so no problem.

I like this project which is along similar lines.

For my own gadget, I need to order stepper motors, think about couplers, and start planning how the electronics will go together. I'm thinking about being lazy and using the parallel port.

I got to see a real RepRap up close!

This evening I went to a presentation and demonstration of a real live RepRap by Bruce Wattendorf and his son. It was very cool to meet somebody who's built a real one and is totally up to speed on every aspect of the project. I asked some questions about the long-term future of the RepRap project.

• Can they get much better spatial resolution without compromising the social goal of serving the developing world? Yes: better spatial resolutions can be gotten with finer nozzles, which would print slower. You could build a duel-nozzle gadget with a wide nozzle for fast clumsy printing, and a narrow fine nozzle for slow elegant finishing.
• Will they bump into patent problems as they move toward the state of the art currently occupied by commercial 3D printers? A number of patents will expire in about three years and the RepRap guys will then be much freer in this area.

He wrapped up his presentation by showing the nanofactory video, "Productive Nanosystems: from Molecules to Superproducts". I came to 3D printers from an interest in nanotech, and he came to nanotech from working on 3D printers. It was gratifying to see that the similarity is clear to people on the other side of the fence.

It was a heck of a lot of fun. I took some pictures. Bruce also has many more pictures on his blog. Interestingly, the parts that are normally plastic in a RepRap are made of wood in Bruce's machine, and he's in the process of printing a set of plastic parts.

Bruce's talk was sponsored by a group called DC401, a bunch of Rhode Island folks who enjoy going to DefCon. They are working with a woman in real estate to arrange a lab space in a building in downtown Providence where they can do electronic and mechanical tinkering. It was fascinating to hear her talk about how she's making it all work by using the other floors for businesses and residential space. This reminds me a lot of MITERS, and it warms my heart.

More developments in cancer treatment

Here are some more new cancer therapies under development. Many of these involve some flavor of nanoparticle (a fancy word for a molecule), and a few involve nanomachines (a molecule that does something more interesting than just sitting there).

• http://www.technologyreview.com/Nanotech/18999/ -- The new nanoengineered system, designed by physician and researcher James Baker and his colleagues at the University of Michigan, contains gold nanoparticles with branching polymers called dendrimers that sprout off the nanoparticle's surface. The particles could be used to launch a multiprong attack against tumors. The dendrimer arms can carry a number of different molecules, including molecules that target cancer cells, fluorescent imaging agents, and drugs that slow down or kill the cells. Once enough of the nanoparticles have gathered inside cancer cells, researchers could kill the tumors by using lasers or infrared light to heat up the gold nestled inside the dendrimers.
• http://www.technologyreview.com/NanoTech/wtr_16690,319,p1.html -- A single treatment of drug-bearing nanoparticles effectively destroys prostate cancer tumors in mice ...the researchers mix together a prostate cancer drug (docetaxel) and polymers that are already FDA-approved... The polymer formed spheres with the drugs trapped within. The researchers then chemically attach pieces of RNA, called aptamers, to the surface of the spheres. The RNA folds into shapes that fit into complementary structures on the surface of prostate-cancer cells... [In placebo groups] almost all the mice died during the experiment. In contrast, all of the mice injected with the targeted nanoparticles survived, and in most cases (five out of seven) the tumors disappeared.
  
• http://www.rsc.org/publishing/journals/CC/article.asp?doi=b800528a -- We present experimental data that demonstrate the potential of synthetic crown ether modified peptide nanostructures to act as selective and efficient chemotherapeutic agents that operate by attacking and destroying cell membranes.
  
• http://www.eurekalert.org/pub_releases/2008-03/uoc--urd033108.php -- Researchers from the Nano Machine Center at the California NanoSystems Institute at UCLA have developed a novel type of nanomachine that can capture and store anticancer drugs inside tiny pores and release them into cancer cells in response to light... the device is the first light-powered nanomachine that operates inside a living cell... [reported on] March 31 in the online edition of the nanoscience journal Small.
  
• http://mednews.wustl.edu/news/page/normal/11449.html -- The nanoparticles are extremely tiny beads of an inert, oily compound that can be coated with a wide variety of active substances. In an article published online in The FASEB Journal, the researchers describe a significant reduction of tumor growth in rabbits that were treated with nanoparticles coated with a fungal toxin called fumagillin. Human clinical trials have shown that fumagillin can be an effective cancer treatment in combination with other anticancer drugs... the nanoparticles' surfaces held molecules designed to stick to proteins found primarily on the cells of growing blood vessels. So the nanoparticles latched on to sites of blood vessel proliferation and released their fumagillin load into blood vessel cells. Fumagillin blocks multiplication of blood vessel cells, so it inhibited tumors from expanding their blood supply and slowed their growth.
  
• http://nano.cancer.gov/news_center/2008/feb/nanotech_news_2008-02-15c.asp -- ...Regulators and drug developers are concerned that these delivery systems may prove difficult to manufacture on a consistent basis... A new study from James Baker, Jr., M.D., PI, Cancer Nanotechnology Platform Partnership at the University of Michigan, and colleagues provides data showing that such concerns can be overcome... the investigators present the results of studies designed to show that they could achieve consistent and specific targeting and cell-killing activity across multiple manufacturing batches of a dendrimer-based therapeutic agent.
  
• http://www.physorg.com/news82653370.html -- A team of investigators has designed a nanoscale, polymeric drug delivery vehicle that when loaded with a widely used anticancer agent cures colon cancer in mice with a single dose... To create their drug delivery vehicle, the investigators used a highly branched polymer, known as a dendrimer, that naturally forms nanoparticles with myriad sites for drug loading. In this particular case, the researchers created what they call a bow-tie polyester dendrimer, whose molecular structure somewhat resembles a bow-tie with two discrete halves... On one half of the dendrimer, the researchers attached a second polymer, poly(ethylene glycol) (PEG), in order to make the dendrimer water soluble... Next, the investigators attached the anticancer drug doxorubicin to the other half of the dendrimer using a chemical linkage designed to break when exposed to acidic conditions. Not coincidentally, the inside of tumor cells is acidic, while the bloodstream has a neutral pH. Results presented in this paper show that the resulting drug-dendrimer formulation releases virtually all of its drug within 48 hours in acidic conditions but less than 10 percent of its payload at neutral pH.
• http://www.azonano.com/news.asp?newsID=4087 -- A new type of cancer detector... the simple and inexpensive system, which can be built from off-the-shelf components, can rapidly detect the presence of cancer biomarkers – telltale proteins in body fluids that can signal the presence of malignant tumors – at very low levels... “With this technology, a future scenario might be that you go to the doctor every year for an annual checkup; he draws about 10 cc’s of your blood and runs it through our machine,” said Soman. “The machine is equipped to detect the biomarkers for all the common types of cancer. Half an hour later it produces a list of the biomarkers that it has found. And then either a software program or the physician examines this list to determine whether you have any cancers that need treating.”
  
• http://nanotechwire.com/news.asp?nid=4703 -- There is a growing recognition among cancer researchers that the most accurate methods for detecting early-stage cancer will require the development of sensitive assays that can identify simultaneously multiple biomarkers associated with malignant cells. Now, using sets of nanoparticles designed to aggregate in response to finding more cancer biomarkers, a team of researchers funded by the Alliance for Nanotechnology in Cancer has developed a multiplexed analytical system that could detect cancer using standard magnetic resonance imaging (MRI).
  
• http://www.forbes.com/claytonchristensen/2008/02/22/cancer-nanotechnology-therapies-lead-clayton-in_jw_0222claytonchristensen_inl.html -- A survey of several different developments, but not much deep discussion of any of them. More of a businessman's-eye view of things, not too surprising for Forbes.
  

TAT variant with magnetic particles

My last posting about targeted alpha therapy discussed the expense of preparing a sample of radioactive actinium, aside from which, targeted alpha therapy should be a very effective and specific and hopefully affordable cancer therapy. Quentin Pankhurst of the London Centre for Nanotechnology has been working with particles of iron oxide, which has very low toxicity and can be attached to antibodies just like the actinium atoms in cages. Iron oxide can be magnetized so each particle can be a permanent magnet. A magnetized particle can then be detected from outside the body using a weak EM field generated by a hand-held device, or it can be heated with a strong EM field, to the point of destroying the cancer cell .

By combining the iron oxide particle with an antibody for the HER2 protein found in breast cancer cells, Pankhurst should be able to achieve the same specificity and effectiveness that Sloan-Kettering has gotten with radioactive actinium, at vastly lesser cost. In order to commercialize this and related applications, Pankhurst has founded Endomagnetics, a start-up based in Houston, Texas.

Why should iron oxide be so much less expensive than radioactive actinium? "Iron oxide" is the chemical name for rusty metal, which is easy to make and store, and readily available in auto scrap yards everywhere. Actinium-225, the isotope used for TAT, has a half-life of ten days, so you can't make a big batch and store some for later use. According to this website at the Oak Ridge National Laboratory: "The actinium-225 is formed from radioactive decay of radium-225, the decay product of thorium-229, which is obtained from decay of uranium-233. The National depository of uranium-233 is at ORNL, and we have developed effective methods for obtaining thorium-229 (half-life 7340 years) as our feed material to routinely obtain actinium-225."

RepRap replicates 100%

This story in Computerworld is a couple weeks old, and I should be working harder to keep up. Vik Olliver, a RepRap hacker in New Zealand (and probably the hardest-working RepRap hacker in the world), has now fabricated all the parts of the RepRap except the Z flag, which can be cut out of the side of a beer can. This includes only the parts that it makes sense to print on a RepRap, so it doesn't include stepper motors, nuts and bolts, pieces of metal and wood (e.g. threaded rods). But it's an important step.

I myself am still drooling a bit over some of the hobbyist CNC stuff. There's a guy in New Mexico who sells these things on eBay. He sells aluminum ones (like this) and ones made of MDF, which I believe is a sort of particle board. Many low-end CNC machines are in the $2000 to $5000 ballpark, whereas he sells these in the $300 to $600 ballpark. It should be pretty easy to swap out that orange router and swap in an extruder.

I was thinking a bit last night about how to drive those steppers, since the offerings on eBay don't include the drive electronics. Digikey sells a stepper motor sequencer chip, the L297, which would be used to drive some power MOSFETs. The L297 just needs an input to choose clockwise or counter-clockwise, and a clock pulse to advance a step in that direction, so you need six GPIO lines to control the three motors, and one more to turn on/off the router or squirt goop out of the extruder. There's some very good information on stepper motors and driver circuits here.

It occurs to me that I've never posted the Sourceforge download page for the RepRap design files. A shocking oversight, given that I want to see the project succeed and proliferate.

Targeted alpha therapy

This is something I read about in 2001, and it still seems to be one of the most promising ideas in cancer therapy. The treatment involves two molecular objects bound together. One is an antibody that gets taken into a cancer cell. The other is a radioactive actinium-255 atom which has a ten-day half-life, and then decays through a few different products, releasing four alpha particles, which rip through the cancer cell and kill it. Luckily alpha particles have only enough energy to destroy one cell, and then they run out of steam and become inert helium nuclei.

At Sloan-Kettering where this work was done, they applied for a patent. A clinical trial was conducted in 2002 with favorable results. There have also been some clinical trials in Australia, I believe.

As far as I am aware, this is a fantastic treatment, due to its being extremely specific, and is applicable to a wide range of cancers, but it's not used much. I would imagine the actinium-255 must be prepared through some process that would probably be very expensive. It would be great if some more affordable alternative could be found. It seems to me that were advanced nanotech available today, some suitable replacement for the radioactive actinium nucleus might be possible.

Nifty stuff over at Machine Phase blog

A couple of interesting things from Tom Moore's Machine Phase blog. One is a comparison between a carbon buckyball and a geometrically similar structure made from DNA using (what appears to be) Paul Rothemund's DNA origami technique. Note the teeny dot in the middle, that's the carbon buckytube.

The other is very interesting because it combines nanotech with my interest in 3d printers in an unexpected way. Specifically it's about using a 3d printer to print parts for an atomic-force microscope, using selective laser sintering. These microscopes typically cost hundreds of thousands of dollars. Hopefully this approach will make them much more affordable for universities, and perhaps high schools and even individual hobbyists.

The white plastic pieces were the things printed with the 3d printer. I always thought of SLS as something done with metal, but I guess it works with plastic too.

My trip to Munich, 20-25 March 2008

My brother Bob and his wife Anja have been spending a year living in Munich. Bob is in the Air Force reserves and is working at the Air Force base there. Here are my photos from the trip.

For the moment this blog entry is only a placeholder, later I'll fill in more details. The itinerary was roughly as follows. Wednesday night, took a red-eye from Boston to London Heathrow, landing at 6 AM local time and spending six hours learning about unreasonable exchange rates. Then a two hour flight to the Munich airport near Erding where Bob lives and works.

View Larger Map
On Friday we went to the Museum Mensch und Natur (map) in downtown Munich. On Saturday we went to Herrenchiemsee, a palace on an island in the middle of a lake, modeled after Versailles in France. On Sunday we went to the Deutsches Museum (photos, map) in Munich, and I've got to say, it's the best science and technology museum I have ever seen in my life. On Monday we went to the castle complex at Burghausen, on the Austrian border. On Tuesday I was too frantic catching my plane for any interesting adventures, although Bob and I did bike into town to get strawberries and fresh croissants.

Nanotube radio antenna work at U.C. Berkeley

Alex Zettl at the University of California at Berkeley has invented an interesting radio antenna made from a single conductive carbon nanotube (less than a micron long and ten nanometers wide) positioned between two conductive plates. He has used the antenna to receive songs transmitted by radio, and has posted the results for your listening pleasure. There is a gap between one plate and a free end of the nanotube, across which electrons tunnel. When a voltage is placed across the two plates, the nanotube's free end becomes electrically charged oppositely from the nearby plate, and the electrostatic attraction keeps the nanotube under mechanical tension.

The nanotube's electrically charged free end moves in response to an ambient radio frequency electric field. This changes the gap size, and therefore the measured tunneling current across the gap, just as with a scanning tunneling microscope. The resonant frequency of the antenna is simply the mechanical resonant frequency of the nanotube under tension. The tension can be changed by changing the voltage across the two conducting plates, and in this way the radio can be tuned. The bandwidth of the antenna is determined by the nanotube's stiffness, and (I think) would depend primarily on the length of the nanotube. The space between the two plates should be a vacuum so the nanotube can move freely, and so that Brownian motion does not detune the radio.

The value of a radio antenna this size is that one can communicate with and control nanorobots, for instance in the human body. One could use these nanorobots for diagnostics, reading out blood chemistry or information about various kinds of cell damage, and could send them instructions to intervene.

There are lots of interesting things happening in the area of nanofabrication, such as Andrew Turberfield's tetrahedra discussed in the previous posting. Presently such things are "controlled" by adding solutions of different DNA sequences to the liquid the structure is sitting in, and the new sequence interacts mechanically with the structure to alter it, by binding selectively with some part of the structure already in place. But each step takes tens of minutes as molecules diffuse through water and position themselves to bind correctly. A signal received by a radio antenna might make things happen much quicker.

Adrian Bowyer interview, Computerworld

Here is the start of the four-page interview. Much of this is stuff that's been published before. Two parts I found interesting appeared on the third page.

Are there plans to modify the current design to replace non-reproducible parts such as bolts with parts that can be manufactured on the machine itself, bringing the overall RepRap design closer to 100 per cent self-reproduction?

Yes - that is definitely one of the evolutionary paths to greater reproductive success. For the immediate future I will be concentrating on widening the list of materials that RepRap can build with (starting with electrical conductors). That widening will implicitly raise the proportion of parts that it can make for itself, of course.

The Fab@Home people have already done a few embedded circuits by printing with conductive silicone. Making circuitry will be a very important ability for these machines.

Can the RepRap recycle what it manufactures?

Yes, recycling has been built in from the start... The main plastic we are using is polylactic acid...

But I want to move to using a non-biodegradable resin. This too is sourced from biomass, but is stable in the ground. That means that the more reprapped goods that get made from it and thrown in landfill, the more carbon is taken out of the atmosphere and locked away for good. And, in 200 years when we have taken so much carbon out of the air to make stuff that anthropogenic global cooling is starting to be a problem, the landfill sites become our strip coal mines to save us.

This is the first time I've ever heard anybody advocate for putting stuff INTO landfills as an environmental measure. An interesting approach to carbon sequestration.