A spokesman for the office of Rep. Eddie Bernice Johnson (D-Texas) says she plans to spotlight the issue of e-voting security, reliability, and integrity at the Democratic National Convention, while an anonymous IT industry source reports that the Kerry campaign is thinking about withdrawing from the Democratic National Committee's endorsement of a voter-verifiable paper ballot requirement in favor of encrypted verification of the final vote tally. However, liberal organizations and some analysts contend that voter turnout could be negatively impacted by the rhetoric surrounding e-voting security and integrity issues, to the detriment of the Democrats. "It's one thing to push for security in all voting machines, but it's another thing to scare people into thinking it's useless for them to go and vote," warns Tanya Clay with People for the American Way. "We can't allow this issue to hijack the election." She says the debate has gotten more intense with a possible conflict of interest when it was revealed that Diebold Election Systems CEO Walden O'Dell donated nearly $10,500 to Republican campaigns between 1997 and 2003, and publicly commented last August that he was "committed" to delivering Ohio's votes to President Bush. Clay also attributes the slow passage of the Help America Vote Act of 2002 to resistance from people who owed their political positions to the outdated electoral process, and explains that e-voting machines are getting the same treatment. Rep. Johnson's spokesman says the congresswoman agrees with Johns Hopkins University professor Avi Rubin that moving forward with the presidential election without resolving e-voting security issues would be "irresponsible."
Click Here to View Full Article

For information regardings ACM's e-voting activities, visit http://www.acm.org/usacm.

Intel's Hans Mulder predicts that wireless communications between sensors and machines will be ubiquitous in two decades, while ON World estimates that sales revenues generated by wireless networks will skyrocket from less than $150 million in 2003 to more than $7 billion by 2010. Driving these predictions is the increasing sophistication of wireless networking technology, and one of the more ambitious concepts is smart dust, super-small sensors that can deployed by the thousands to monitor forests for fire, battlefields for hazardous materials, and borders for signs of intrusion, to name just a few examples. The smart dust concept envisions sensors equipped with antennas, radio chips, and a power source or battery that are as small as a grain of sand; this is a formidable challenge, but postage stamp-sized sensors with lithium batteries could become a reality in a few years thanks to breakthroughs in mesh networking, in which sensors only need enough power to chat with their neighbors, which pass information along to a central computer or control station. Wireless sensor networks can be set up quickly and easily reconfigured. Such networks would watchdog areas or situations where human monitors or wired system monitoring is impractical. Sensor networks can not only function as data collectors, but as automated control systems for other machines: An example is the Defense Advanced Research Projects Agency's use of Sensoria technology to build mobile antitank landmines that move to fill in the spaces left by exploded mines. Dust Networks CTO Kristofer Pister thinks the size of sensors is a less important issue than "reliability, power and cost," while signal interference and security also need to be addressed. Thus far, most mesh systems have performed less reliably in the field than in controlled laboratory settings, while one potential power-saving strategy is to leave sensors in a sleep mode most of the time.
Click Here to View Full Article
(Articles published within 7 days can be accessed free of charge on this site. After 7 days, a pay-per-article option is available. First-time visitors will need to register.)

Xerox's Palo Alto Research Center (PARC) has entered into a partnership with the nonprofit Scripps Research Institute to develop new technologies that could revolutionize medicine and medical research. One tool yielded by the joint effort is the nanocalorimeter, a device that can detect super-small changes in temperature caused by the interaction between a drug and its target, and that could be used by pharmaceutical companies to help them rapidly select the best drug candidates and bring improved drugs to market faster. Scripps President Dr. Richard A. Lerner says the biomedical solutions PARC is developing are a natural extension of Xerox technology: For instance, the nanocalorimeter's function depends on the precise deposition of drops of liquid on the surface of the device, which essentially is what photocopying does. Another PARC-Scripps biomedical tool announced earlier this year is the Fiber Array Scanning Technology cytometer, a laser scanning device that can theoretically detect one cancer cell out of 10 million healthy cells in a blood sample. Lerner says the device could enable earlier cancer diagnosis as well as quickly measure the effectiveness of chemotherapy. PARC computer science laboratory manager Richard Bruce notes that the cytometer finds cells using built-in laser printing scanning optics. The products that come out of the partnership will not be commercialized by either PARC or Scripps--instead, the technology will be licensed to large companies or spun off into marketable inventions. Other areas of development for the partners include sutures equipped with diodes that detect cancer cells and kill them with ultraviolet emissions as they sew up the perimeter of a removed tumor, as well as analysis of changing protein patterns in cells in order to determine why transplanted organs are rejected. Bruce says PARC's work with Scripps grew out of its desire to work on "hard problems."
Click Here to View Full Article

Political parties have gotten a lot of mileage out of blogs, candidate home pages, and Web-based fundraising when it comes to linking candidates with their core support base--but as the presidential election approaches, personal digital assistants (PDAs) and databases are coming to the fore as Democrats and Republicans vie for swing-state voters. Volunteers will canvass these states with PDAs or clipboards over the coming months, interviewing residents and sending the information they glean to political databases such as the Republican National Committee's Voter Vault. The databases will then be used to identify supporters and ensure that they vote on election day, as well as find unregistered voters and discover what undecided voters' interests are in order to tailor campaign material to them. The databases are being used to track nearly 200 million names, presenting a daunting data processing challenge, since much of the data comes from local voter registrar offices located all around the country and the data must be "cleaned." Such activities were once done by telephone, but public irritation at telemarketers has lessened the effectiveness of such an approach. There are concerns among many Democrats that the Republicans have better technology, mostly because they started their database project over 10 years ago. However, PlusThree's David Brunton thinks Democratic efforts have narrowed the database chasm between the two parties. Meanwhile, the more sophisticated databases have enabled the parties to sort the data for more particular information, similar to the ways companies do for marketing purposes, such as whether someone prefers dogs to cats. However, the uses and value for such data is still largely untested.

Hewlett-Packard researcher Bernardo Huberman is conducting research on how information flows within an organization in order to understand how those flows can be harnessed to predict business events. Huberman's research team studied HP Labs' global email network for three months and used a series of test emails with embedded URLs to track information flows; the group discovered informal social structures and communities of interest that generally placed higher-level managers near the centers of information exchange. The study also found that information seldom went two nodes beyond the original recipient. Huberman believes the research will help use information flows within an organization to identify real information leaders as well as to create more accurate predictive business analysis. Traditional methods of plotting future business action usually involve a series of meetings after which a key group participant takes responsibility for acting on the consensus outcome. Using informal information flows weighted according to financial market theory, Huberman's approach aggregates information from the best sources within an organization to provide better predictive results. The research study validated its theory by offering a group of financial investors cash incentives to correctly predict stock performance, using tools to control peer pressure and estimate the attitude of predictors; the aggregate forecasts of the group, weighted in favor of the best performers, provided results superior to that of any individual investor. Hewlett-Packard's services division applied Huberman's information-flow algorithms to predict its September revenues, and the company also now uses the technique to predict monthly DRAM memory prices.
Click Here to View Full Article

Carnegie Mellon University's Grace and George robots are tackling the Open Interaction Task at this week's Mobile Robot Competition and Exhibition, sponsored by the American Association for Artificial Intelligence (AAAI) in San Jose, Calif. The robots, developed in collaboration with researchers from the Naval Research Laboratory and Swarthmore College, will work in conjunction to interact with conference attendees throughout the duration of the meeting. Grace will provide conference and schedule information at a booth while George will mingle with conference-goers and guide people to various conference locations. Grace will also make appointments for George to meet attendees at certain times and lead them to where they need to go. Attendees to meet and be led by George must wear a specially colored hat. Grace competed in the 2002 Robot Challenge where she successfully found the registration booth, registered, made her way to the conference room, and delivered a PowerPoint presentation from the podium. Carnegie Mellon Robotics Institute research professor and project coordinator Reid Simmons says the Open Interaction Task better showcases the robots' ability to interact with humans. The project goals include advanced human-robot interaction, more robust voice recognition, and to eventually have robots replace conference volunteer workers. More information on Grace and George is available at www.ri.cmu.edu/projects/project_522.html.
Click Here to View Full Article

The purpose of the Defense Advanced Research Projects Agency-funded Strong Angel II project is to set up a communications system for efficient collaboration between military and civilian personnel to enhance the coordination of disaster relief efforts. Dan Gillmor reports that a recent test of the system on a lava bed in Hawaii was not only significant from a technological point of view, but from a collaborative perspective, as it united people with diverse, often antagonistic, political views. Technologies employed in Strong Angel included laptops running with collaboration software from Groove Networks; Web cameras tested by VSee Lab founder Milton Chen; and both proprietary and open-source software. One demonstration involved a vehicle equipped with wireless networking and laptops that collected and disseminated data as it passed from location to location. Such an approach can enable humanitarian workers to gather data in the field without setting up complex communications systems in disaster areas. Another demonstration involved the capture and recording of Arabic news broadcasts and the translation of their audio and text into English by software. The broadcast and the Arabic and English texts were placed in a database, and a member of BBN Technologies Speech and Language Processing Group extracted the most significant video segments and put them in a file that was sent to a human translator on the U.S. mainland. Such technology would be of enormous help for U.S. forces and humanitarian workers in Iraq, where fast knowledge of media opinions about the war and the American occupation are of great import.
Click Here to View Full Article

A patent has been awarded to a team of Toyota inventors in Japan for an automobile that can reproduce facial expressions to communicate the driver's moods to other drivers. The inventors write in the patent that "as traffic grows heavier and vehicle use increases, vehicles having expression functions, such as crying and laughing...could create a joyful, organic atmosphere rather than the simple comings and goings of inorganic vehicles." They add that the driving experience would be more comfortable for occupants that have a close rapport with their cars. The patent describes a car with a wagging antenna, adjustable body height, headlights with variable intensity, and hood features that resemble eyebrows, eyelids, and tears, all of which can glow with colored lights chosen because of their psychological value. The car is equipped with a computer and software system that measures road and vehicle conditions, while drivers and passengers can also enter data about their emotional states. The computer collects and stores this data, accumulating points attributed to certain factors--an occupant's reaction to someone cutting in front recklessly, sudden braking because of an unexpected pedestrian, mechanical difficulties, etc.--and the software triggers an appropriate emotional display once a certain threshold has been reached. The patent describes an indication of anger as red hood lights and illuminated eyebrows, and an indication of happiness as a vibrating antenna and orange hood and eyebrow lights that shade the headlight so it appears to wink. Surprise is relayed by orange hood lights, red eyebrows, shaded headlights, and lowered vehicle height in the rear.
Click Here to View Full Article
(Articles published within 7 days can be accessed free of charge on this site. After 7 days, a pay-per-article option is available. First-time visitors will need to register.)

The federal government was urged to make a greater commitment to cybersecurity and to have cyberspace experts take on a larger role in Homeland Security efforts during a hearing before the House Science Committee on July 21. Cybersecurity experts said more educational programs are needed, and added that courses will have to be up-to-date and be able to adapt to the latest demands of cyberspace. Chet Hosmer, president of Wetstone Technologies, a cybersecurity research development company, stressed that security experts will need to make adjustments quickly because potential attacks can develop and change at "Internet speed." Hosmer also took issue with the setup of higher-education curricula, which is producing fragmented cybersecurity training programs because of its rigidity. The social science department offers criminal justice programs, while computer science is relegated to math or computer science departments. "Building programs that cross domains is quite difficult for many reasons, and the student typically lacks depth in either area and is ill-prepared for [work in] digital investigation after graduation," said Hosmer. Some educators saw community colleges as an ideal resource for security training because of their focus on practical skill. Military educational programs, such as the National Strategy to Secure Cyberspace, are another form of cybersecurity training.
Click Here to View Full Article

Technologists from Germany's Fraunhofer Institute for Media Technology this month showcased their new 3D audio technology to Hollywood movie studios. MP3 co-developer Karlheinz Brandenburg led a team to develop the new audio technique, which requires abundant processing power, a sophisticated mathematical algorithm, and an array of up to 400 small speakers. The effect of the 3D audio technology, dubbed losono, is impressive, according to those who have experienced it: Sound engineers would be able to create sounds that sound as though they are coming from a precise location. Current stereo technology has advanced to the point where it can create ultra-realistic sound for listeners positioned in an area just a few feet wide, but losono would enable sound designers to create an audio illusion for everyone in an entire room. Brandenburg says processing power is a critical element that enables losono to work, because the Fraunhofer Institute's algorithm works to distribute the bits and pieces of a sound to the vast array of small speakers in order to create the desired effect. Fraunhofer predicts the 3D audio technology will first be deployed in theme parks and other presentation spaces, but eventually will migrate to movie theaters and home audio systems as prices go down and content is developed. Losono requires special equipment for sound designers and mixers, including a console with a light pen that lets creators place sounds in an environment; more specialized equipment at the theater or other venue would translate those instructions back into a configuration that fits the particular room size and shape. Even though losono is expensive and complex, experts who have experienced the technology predict it will be adopted simply because of its unique capabilities.
Click Here to View Full Article

Asianux says it will get its first South Korean endorsement soon, a key development if its server operating environment is to become a Linux Standard in Asia. At the Oracle OpenWorld conference in Shanghai, Chris Zhao, acting president of China's Red Flag Software, said he expects a Korean partner to come aboard within two months. Asianux 1.0, unveiled in July, is the work of Red Flag and Japan's Miracle Linux, which view support from the other North Asian hotspot for open-source software as being crucial to capturing the necessary backing of Linux distributors throughout Asia. The collaboration across the three markets comes as government officials have expressed interest in using an open-source alternative to Microsoft Windows, and the companies have put the server operating environment in a good position to fulfill such a roll. Asianux has attracted more than 40 hardware and application vendors so far. "Now that fragmentation has been removed, we have a standard, certified Linux version we can count on and that's going to mark for big change for this area of the world," says Charles Phillips, co-president of Oracle, which owns a majority stake in Miracle Linux.
Click Here to View Full Article

The purpose of the Information Society Technologies program-funded SECRETS project was to assess the advantages and disadvantages of open source software for Internet security for the benefit of the public and private sectors, and its evaluation of the toolkit for deploying OpenSSL's Secure Sockets Layer (SSL) and IPSec's Free Secure Wide Area Network (FreeS/WAN) yielded mixed results. The protocols' functionality was tested in the areas of secure e-commerce, secure mobile communications, network monitoring, and intelligent networks. Intrasoft International's Antonis Ramfos reports that one of open source software's major drawbacks is that the organizations that devise the protocols frequently fail to capably support them afterwards, while a dearth of standardization has led to interoperability problems with other open source software. Such problems were typical of FreeS/WAN, according to the SECRETS evaluation. Despite such problems, Motorola's Ross Velentzas says the SECRETS project determined that the protocols' deployment is "worth considering by commercial organizations and governments for integration into the software products" they build or employ. The utilization of OpenSSL by others is much easier than FreeS/WAN because, unlike Free/SWAN, OpenSSL boasts sufficient documentation from its organization. The SECRETS partners, which include Motorola, Intrasoft, and Alcatel, are still working with open source software for Internet security, and Ramfos and Velentzas concur that both the public and private sectors will use such protocols more extensively in the future.
Click Here to View Full Article

Both the Democratic and Republican rivals for the presidency have made IT's role in business, the economy, and the upcoming election abundantly clear. Democratic candidate John Kerry and President Bush agree that U.S. companies should have a global presence, while broader and more affordable broadband access is needed for IT workers to maintain their productivity. Cybersecurity needs to be improved, and personal data must have guaranteed privacy; Bush and Kerry also think that health care services can be made more efficient and less costly through IT, and both favor the preservation of intellectual property's value and fulfilling the potential of emerging technologies such as nanotechnology. But the candidates have strikingly different strategies for dealing with the hiring of offshore labor, the delivery of broadband access to rural and inner-city areas, and how fast digital medical records should be made widely available without compromising personal privacy. Kerry's stance is to remove tax rewards for companies that offshore jobs, while Bush favors leaving businesses alone in order to avoid "economic isolationism;" however, UC-Davis professor Norman Matloff argues that both Democrats and Republicans "are solidly pro-H-1B and pro-offshoring." Kerry wants to set up tax incentives for companies that deliver broadband access to underserved areas and invest in faster networks, whereas Bush wishes to make a moratorium on Internet-access tax permanent. Bush has stressed broadband's value to U.S. security in particular. Kerry aims to give "every" American an electronic health record within four years--a goal that Universal Health CIO Linda Reino calls unrealistic--while Bush wants e-health records for "most" Americans within a decade, and has named a national health IT coordinator to see it through.
Click Here to View Full Article

Virginia Polytechnic Institute and State University hosts System X, the third most powerful supercomputer in the world according to Top500.org. System X cost $10 million to build and install, a mere fraction of other leading supercomputing projects thanks to its use of 1,100 commercially available Apple Power Mac dual-processor G5s, which were recently phased out in favor of less power-consumptive Apple Xserve units. Virginia Tech computer science professor Srinidhi Varadarajan explains that simulation alone is inadequate when it comes to testing new programming models or memory management methods on very-large-scale systems. He says thermal heat was a major issue with System X, and plans to use IBM Opteron processors were dashed because they would need two dual-Opteron machines to perform both multiplication and addition operations, which made the cost prohibitive. But Apple's Unix-based PowerPC 970 platform suited their needs perfectly. System X's initial benchmark results were less than stellar, but they picked up as time went on: Only 800 billion floating-point operations per second (FLOPS) were recorded on Oct. 1, but 10 days later System X was at 7 trillion FLOPS, with an additional 10 percent being churned out every day. Varadarajan notes that his team developed operating system-independent and hardware-independent management software to increase System X's fault tolerance, which is currently being commercialized. The professor says that System X's success bucked the odds, as the project involved an untested communications infrastructure and cooling system design, as well as commercial processors that had never before been employed in such a large project.
Click Here to View Full Article

As part of its push to fully realize network-centric operations, the U.S. Department of Defense is calling for more sophisticated sensor technology to support persistent surveillance and other operations to ensure fewer casualties and more efficient military operations. Achieving this goal requires integrators to devise effective techniques for networking sensors and managing the data they generate, although experts concur that technology alone cannot address net-centricity. Training the military to use the new systems and not devaluing human intelligence must also be key concerns. Boeing's Carl O'Berry believes self-organizing, mobile ad hoc sensor networks will guarantee that information is sent to the correct command center as well as facilitate communications in bandwidth-scarce environments. Experts say that network-centric operations in both times of peace and times of war will have a heavy reliance on data sharing, visualization, and analysis applications; one project along these lines is Batelle Memorial Institute's Starlight, a visualization tool employed in Iraq that outlines a 3D model or digital map of trends and relationships by pulling in large sets of XML data. Managing the tsunami of information produced by sensors and making it network-accessible is the goal of the Defense Information System Agency's (DISA) Network-Centric Enterprise Services (NCES) initiative, which will consist of nine core services comprising the software infrastructure on which network-centric applications will operate. DISA believes NCES will be developed by integrators in a "living lab" where the community can access source code and where mature prototype services can be tested in operational scenarios. There is concern among critics of network-centric warfare that such projects could lead to information overload, but BAE Systems' John Osterholz is confident that this issue will be resolved by the new application standard.

IBM Almaden researcher T.V. Raman says multimodal computer applications are gaining ground with the convergence of best-of-breed voice and visual language standards. Raman focuses on integrating speech into Web applications and is one of the inventors of the XHTML+VoiceXML (X+V) language, the goal of which is to bring together W3C standards and integrate them using a common event model, so that voice input affects the graphical user interface (GUI) and vice versa. X+V language uses XML Events for enabling VoiceXML dialogs to function as event handlers and create XMTML-based results in the GUI, for example. Multimodal applications that make use of speech are especially attractive for in-vehicle computer applications, personal digital assistants, home-entertainment systems, and other applications where there is already a display and voice capability is a logical next step. Raman admits voice applications have not caught up with expectations yet, but he says there are still many important uses for voice applications today: VoiceXML used in voice-only applications is an important first step that will enable users to incrementally test multimodal X+V applications at the same time. Raman says multimodal and voice-only applications do not compete with each other. Europe is a likely starting point for many future multimodal efforts because of its advanced cellular infrastructure, while there is a tremendous Asian market because of the current difficulty in using keyboards in conjunction with character-based languages. In developing nations, multimodal PCs will enable illiterate people to use computers, as is happening in rural Indian villages with the Simputer.

Project Oxygen is a five-year, $50 million interdisciplinary program managed by MIT's Computer Science and Artificial Intelligence Laboratory to design pervasive computer systems that people can easily communicate with, and that can be seamlessly woven into households and workplaces to the degree that they become a standard component of daily life. Intelligent workspaces that automatically adapt themselves to their users' routines, computer chips that customize themselves to different applications, and location-aware sensors for personal navigation are just some of the working prototypes being turned out in Project Oxygen's fourth year. Project Oxygen participants envision a future where mouse-and-keyboard interfaces can be replaced by software that comprehends spoken requests and advanced vision systems, thus enabling remote-control computer interaction. Project Oxygen co-director Victor Zue says the program is focusing on the combination of networking, speech, and vision technologies into interfaces that can be controlled by voice as well as gestures, while other projects are exploring the development of computers that can anticipate users' needs by studying their behavioral patterns. A group led by principal research scientist Howie Shrobe is developing intelligent spaces that respond to vocal commands, and is working on linking these areas to a software platform that will enable physically scattered persons to exchange and display information with whatever devices are on hand. Another prototype system is Cricket, a wireless location-aware technology that involves wall- and ceiling-mounted beacons that work with a receiver carried by a person to locate that person to within a few centimeters. Some Oxygen projects are working to address potential questions about first-generation pervasive computers, such as how secure they are from hacking.
Click Here to View Full Article

Sun Microsystems fellow James Gosling, who is credited with creating the original design of Java and the Java Virtual Machine, puts security issues with virtual environments into perspective by stating in an interview with Sendmail founder and CTO Eric Allman that security problems are the result of poor implementation rather than an elementary flaw of the environments themselves. In regards to the kind of overhead a person should anticipate in an virtual environment, Gosling explains that "When you talk about a virtual machine, that doesn't necessarily imply that it's actually interpreted. You can go through some kind of transformational step and actually end up with a machine code that is what is actually executing." He points out that performance is not a major concern with many dynamic languages, and says that raising Java's performance, security, and reliability to a plausible stratum was a design objective from the start. Gosling thinks Microsoft was foolish in deploying C# as a virtual machine environment that supports languages such as C and C++, which essentially throws out security. "Security is one of those things that you don't add by painting it on afterward," he remarks. Gosling thinks that Java's mantra, "write once run everywhere," is holding up well, but becomes a little shaky in situations where applications are highly dependent on environmental specifics, such as the size of the display.