Dr. Dobb's Journal June 2000
The World Wide Web may seem chaotic, but it's tame in comparison to what's in store as "Freenet" becomes available over the next few months. Based on the college thesis of Irish computer-science student Ian Clarke, Freenet has become a volunteer project to make the Internet completely free of centralized control. Clarke's thesis, "A Distributed Decentralized Information Storage and Retrieval System," is available at http://freenet .sourceforge.net/.
Freenet has no centralized servers, Uniform Resource Locators (URLs), or Domain Name System (DNS). Information is stored randomly (geographic proximity is the only initial criterion) in caches of the computers (nodes) connected to Freenet. Each data item stored on Freenet is accessed by a unique key, but the exact location of the data becomes unknown as it is passed from one node to another. Consequently, it will be impossible to impose censorship, copyrights, or any other form of control on the information stored on Freenet.
Freenet is not the first peer-to-peer distributed network on the Internet. Gnutella (http://gnutella.nerdherd.net/) is another, but currently only runs on Windows. Since Freenet is based on Java, it is more likely to gain broad acceptance. To become a Freenet node, you need the Freenet Daemon (written in Java), which is available from the Freenet web site. As DDJ goes to press, Freenet is nearing a public release as a basic prototype.
Nanoengineering holds promise for advances in medical treatment and diagnostics, data storage, and other applications. But perhaps the biggest challenge has been in creating molecules that have consistent magnetic properties, which are key to the precise control and reliable operation of nanodevices -- and researchers from several labs have recently announced breakthroughs in this area.
IBM researchers have developed "self-assembling" materials by applying chemical reactions (a mixture of iron and platinum) that cause magnetic nanoparticles to automatically arrange themselves into well-ordered arrays, with each particle separated from its neighbor by the same fixed distance. They think this breakthrough could result in data storage densities up to 100 times the current record (held by IBM) of about 4 gigabits per square centimeter.
Meanwhile, an Indiana University project, led by chemistry professor George Christou, has produced single-molecule magnets (SMMs) from manganese, which function as magnets at low temperatures. The trick, according to Christou, is to find SMMs that function at higher temperatures. SMMs could lead to storage densities of up to 30 terabits per square centimeter, almost 10,000 times the current IBM record.
Finally, a Georgia Tech team has been working on methods for dynamically controlling magnetic properties of nanoparticles to be used for medical treatment in the human body. While storage devices require stable magnetic properties, particles used to deliver drugs to targeted areas in the body, or for medical imaging, need to have properties that allow the magnetic state to be varied constantly. For more information, see http://www.acs.org/meetings/sanfran2000/.
Researchers at Lawrence Berkeley National Lab (http://www.lbl.gov/) have produced for the first time light pulses lasting less than 300 millionths of a billionth of a second -- 300 femtoseconds, in other words. The spectral range of these pulses, which were produced off the primary beam of a synchrotron light source (the Advanced Light Source electron synchrotron), extend from infrared to X-ray wavelengths. The researchers expect X-ray pulses of 100 femtoseconds to be produced in the near future. What's important about this breakthrough is that it will let scientists capture the motion of atoms during physical, chemical, and biological reactions on an infinitesimally small time scale.
A DDJ-sponsored team of computer- science students from the California Institute of Technology (Caltech) outscored all other U.S. programming teams and came in 11th overall in the Association of Computing Machinery (ACM) International World Programming Contest World Finals. Sponsored by IBM, the annual programming contest, now in its 24th year, challenges teams of college computer-science students to solve difficult programming problems. Initially, more than 2400 teams from all over the world entered the contest, which was narrowed down to 60 for the World Finals. The winning team was from St. Petersburg, Russia. In the top ten, two teams were from Russia, three from China, two from Canada, and one each from Australia, Japan, and Germany. The Caltech team, which consisted of Benjamin Mathews, Christopher Chang, and Miroslav Dudik (and coached by David Epstein) tied for 11th with Bangladesh University, Charles University of the Czech Republic, and ZhongShan University of China. For more information (including the programming problems), see http://acm.baylor.edu/acmipc/ WhatsNew.htm.
The Israel Institute of Technology (Technion) has developed a robot for performing knee replacement surgeries with greater precision and less error than human surgeons. The Technion Robot has a steadier hand than human surgeons, according to its developer, mechanical engineering professor Moshe Shoham. Orthopedic procedures involving rigid bone are well suited for robotic surgery because precise coordinates can be gleaned from computerized tomography scans. According to Shoham, about 10 percent of the 200,000 knee replacement surgeries done in the U.S. each year must be repeated due to surgical error. Shoham sees a future for surgical robots in other surgical procedures such as spinal and eye surgery. See http://www.ats.org/v2/ for more information.
The University of California at Berkeley's department of electrical engineering distributes a software model for designing integrated circuits that has become widely adopted by semiconductor manufacturers. Called the "Berkeley Short-channel IGFET Model" (or BSIM), it can model the operation of virtually any number of transistors linked in an integrated circuit. In a nutshell, BSIM is simply a mathematical equation that simulates the operation of a transistor. A new version of the model, BSIM4, can model transistor operation in the time scale of one hundredth of a trillionth of a second and has new functions for modeling noise. BSIM4 is freely available at http://www-device.EECS.Berkeley.EDU/ ~bsim3/bsim4.html.