President Bush's call for doubling the federal funding of basic scientific research comes as a response to several meetings that White House officials held with technology executives and educators. Bush's plan to request $910 million in the first year, and $50 billion over 10 years was welcome news for computer scientists who have long warned against the destructive impact of eroding federal funding. Bush identified nanotechnology, supercomputing, and alternative energy sources as long-term initiatives that the administration would now support, in a departure from its traditional focus on short-term research. Under Bush's plan, spending would likely increase by 7 percent annually, roughly doubling over 10 years. While the details remain vague, ACM President David Patterson is excited: "This is really a huge deal and I'm very encouraged," though he noted with concern that many legislators attending the State of the Union address were not moved to applause by Bush's announcement. "It just shows the challenge we have." In two high-profile discussions where the administration was urged to heed the warning of the National Academy of Sciences that science and technology education are eroding rapidly, Intel Chairman Craig Barrett met with Vice President Dick Cheney, and Charles Vest, the former president of MIT, met with OMB director Joshua Bolten. The executives and educators who had attended those meetings were still unsure if the administration would act on their recommendations, so Bush's announcement came as a welcome surprise. "We haven't seen this interest in basic research from this president before," said the American Association for the Advancement of Science's Albert Teich. The growing problem of funding for research and education has also attracted Congressional interest, as two bipartisan bills addressing the matter have recently been introduced.
Click Here to View Full Article
to the top

Data Mining Tells Government and Business a Lot About You
Knight-Ridder Wire Services (02/01/06) Boyd, Robert S.

Data mining is at the center of the debate over the government's warrantless eavesdropping program, as the same complex algorithms that identify consumer preferences and search tendencies are being used to ferret out suspicious relationships that could point to terrorist activity within the nation's borders. The Senate Judiciary Committee will open its investigations into the government's eavesdropping program on Monday. Data mining is dependent on computers to cull through an ocean of meaningless data to arrive at one potentially meaningful relationship, at which point a human investigator takes over. Determining whether two members of a list of 10 suspects have ever stayed in the same hotel at the same time would require a search of at least 250,000 records, well beyond the capacity of a human, notes Stanford University computer science professor Jeffrey Ullman. "Before data aggregation and data mining came into use, personal information on individuals contained in paper records stored at widely dispersed locations, such as courthouses or other government offices, was relatively difficult to gather and analyze," according to a Congressional report issued by the Government Accountability Office (GAO) last year. The GAO estimates that 52 government agencies had implemented or were planning data mining initiatives in 2004. Should a data mining initiative be employed on a national scale, it might include a search engine that would scan for suspicious words, such as "bin Laden" or "nuclear plant" in telephone and email communications. Data visualization techniques could plot out these words, giving due weight to their frequency. Data mining is also widely used in the insurance industry to detect fraud, as well as in marketing, politics, and finance. Increased data mining usage has naturally drawn criticism from privacy advocates, who argue that while the program cloaks itself in national security concerns, it nonetheless is overly invasive.
Click Here to View Full Article
to the top

U.N. Tech Summit Promotes Middle East
TechWeb (01/31/06) Jones, K.C.

This April, Oman will host the first World Summit on Innovation and Entrepreneurship, which will bring together more than 1,000 leaders in education, business, science, and policy to promote technology and innovation in the Middle East. The summit, which plans to develop a workable model that can be applied to other emerging markets such as Africa and Eastern Europe, will draw representatives from Microsoft, Google, MIT, Hewlett-Packard and many other organizations. One summit participant, Cisco, has already made significant inroads in the Middle East, having established all-women's technical academies in countries such as Jordan and Saudi Arabia. Oman was selected to host the summit for the progressive policies of its monarchy, which has pursued a program of "Omanization" to deliver education and job opportunities to its citizens. Global Leadership Team Chairman Sam Hamdan said the summit would focus on empowering developing societies with technology and overcoming the digital divide. The summit will address six principle issues: improving existing infrastructure; accelerating knowledge partnerships through empowerment; inspiring human talent to sustain development; legal reforms and improved governance through innovation; information sharing to spur growth; and examining the effect of innovation on future generations. Among the other issues on the summit's agenda are bringing women and children into the technology constellation and a host of environmental and geopolitical issues, such as renewable resources, managing oil wealth, and intellectual property rights. Technological topics will include nanotechnology, e-government, telecommunications, and open source applications.
Click Here to View Full Article
to the top

The Open-Source Programmer Who Means Business
CNet (02/02/06) Marson, Ingrid

In a recent interview, Red Hat's Alan Cox outlined his thoughts on a host of open-source issues, including the Linux kernel, software patents, and the new version of the GPL. Cox applauds the GPL for its compatibility, and describes the update as a more sensible version of GPL 2.0. In the wake of the Sony rootkit scandal, Cox believes that digital rights management may ultimately be determined by government regulation, as a consensus has yet to emerge on the relationship of computers and private property. An ardent opponent of software patents, Cox expresses shock that the European Commission would consider altering the patent process after the European Parliament flatly rejected software patents last year. Meanwhile, pressure from IBM and Microsoft to reform the U.S. patent system is progressing slowly, but Cox believes that it is evolving in the right direction. He is pleased with the development of the Linux kernel, noting that it has evolved to a point where it offers all the functionality that its users require, and that the modifications and updates that are currently being issued are mostly tweaks--subtle improvements on speed and efficiency. Cox praises the unstructured development process of the Linux kernel, though he notes that every change is reviewed, and that there are few prominent independent developers anymore. Migration to Linux is gaining steam, particularly in larger operations and environments where PCs are used mostly for word processing. Cox also notes that OpenOffice has been a key driver of Linux adoption.
Click Here to View Full Article
to the top

U.S. High-Tech Jobs Waking From Long Slumber
Investor's Business Daily (02/02/06) P. A4; Much, Marilyn

The productivity gains of the U.S. technology sector is prompting the industry to hire more tech workers, according to Mark Zandi, chief economist of Moody's Economy.com. Job growth rose to 125,000 positions last year, up about nine times from 14,000 in 2004, says Zandi, who expects high-tech companies to add 217,000 jobs this year. Going forward, he anticipates a significant run up in the demand for tech products globally over the next few years, prompting him to forecast the creation of 126,000 jobs in 2007, 123,000 in 2008, and 150,000 or more a year in 2009 and 2010. Meanwhile, the creation of 2,000 tech jobs in Silicon Valley in 2005 marks the first time the region added new jobs in four years, according to the trade group Joint Venture: Silicon Valley Network. Russell Hancock, chief executive of the organization, says most of the new tech jobs in the region were high-end positions, and that growth is likely to continue in areas such as biotechnology and nanotechnology. In another good sign, outplacement consulting firm Challenger, Gray, & Christmas reports that the number of tech layoffs fell to 174,744 in 2005, down from 176,113 the previous year. However, the firm believes mergers and acquisitions, including private equity buyouts, will provide an element of uncertainty in the years to come.
Click Here to View Full Article
to the top

Georgia Tech, Oak Ridge and UT-Battelle Collaborate on High-Performance Computing
EurekAlert (02/01/06)

Large-scale research using supercomputing technology in the United States should receive an enormous boost as a result of a collaboration between the College of Computing at Georgia Tech (COC), Oak Ridge National Laboratory (ORNL), and UT-Battelle. The partnership calls for the research bodies to share their facilities, scientific resources, and staff. COC is a leader in using computer technology to advance social and scientific research, ORNL is the largest multipurpose laboratory of the U.S. Department of Energy, and UT-Battelle is a nonprofit partnership between the University of Tennessee and Battelle that manages ORNL. Dr. Thomas Zacharia, associate laboratory director for ORNL's Computing and Computational Sciences Directorate, is set to become a professor at COC. In the months to come, there will be other appointments involving faculty and staff members, and computer resources will continue to be shared. Also, COC's Computational Science and Engineering Division will open a campus at ORNL. "We firmly believe that this partnership with ORNL and UT-Battelle will create a one-of-a-kind environment for high-performance computing research and help reinvigorate U.S. capabilities in supercomputing," says Richard A. DeMillo, dean of the College of Computing at Georgia Tech.
Click Here to View Full Article
to the top

Spyware Poses a Significant Threat on the Net, According to New Study
UW News (02/02/06)

A recent University of Washington study of more than 20 million Internet addresses has found that spyware, while slightly less prevalent today than last spring, is nonetheless alive and well. The researchers found that more than one in 20 executable files carried spyware, while one in 62 domains attempt to forcibly upload spyware for users who merely visit a site. Piggybacked spyware appeared most frequently on celebrity and game Web sites. "For unsuspecting users, spyware has become the most 'popular' download on the Internet," said computer science and engineering professor Hank Levy. Spyware severity runs a wide spectrum, as some programs simply inundate users with popups, while others can steal personal information, install malicious programs without the user's knowing, or even corrupt the computer beyond use. Of the piggybacked spyware that the researchers found, 14 percent carried potentially malicious applications, while the rest was relatively harmless adware. Drive-by download attacks were down 93 percent from May to October of last year, which could be explained by the increasing use of antivirus software and automatic update programs. The researchers recommend that users have at least one anti-spyware program installed on their computers and that they keep it up-to-date, and that users exercise common sense and only download from reputable sites.
Click Here to View Full Article
to the top

High Tech, Under the Skin
New York Times (02/02/06) P. E1; Bahney, Anna
(Access to this site is free; however, first-time visitors must register.)

The convergence of man and machine is taking a bold step forward as people implant RFID devices under their skin to log on to their computers, unlock their cars, and open doors with a wave of the hand. Some enthusiasts argue that the cell phone has essentially become an appendage of the human body, and implanting a chip is simply the logical extension of our intertwined relationship with technology. RFID tags have been implanted in livestock for years, enabling owners to scan the animal from two to four inches away to determine if it belongs to them. While the blending of humans and computers has long been the dystopic fantasy of science fiction visionaries, a growing body of pro-convergence technologists see its reality in the redesigns of cameras, MP3 players, and storage devices to resemble jewelry and blend into a user's wardrobe, as well as the jackets and sunglasses that now come with Bluetooth capabilities, enabling them to function as digital devices. For the many people who see the cell phone as an extension of themselves and carry flash drives on their key chains and iPods with their entire music library, an implanted RFID chip is no different than having a filling put in, says the Institute for the Future's Alex Soojun-Kim Pang. RFID tags can be obtained on the Internet for as little as $2, and devices such as computers and car doors can be modified with wire scanners to interface with the chip. Implanting RFID chips is not a new practice, as the Florida company Verichip has implanted more than 2,000 people with chips to link to their medical records since 2004. The practice has drawn criticism from privacy advocates, who claim that while the technology may be an appealing novelty today, the future could bring an RFID-dependent climate where people will be required to have the chips implanted. There are also health concerns about the procedure, as many have had their chips implanted in non-medical settings.
Click Here to View Full Article
to the top

PC Industry Looks to Transform Firmware
eWeek (01/31/06) Spooner, John G.

The computer industry is moving toward the United Extensible Firmware Interface (UEFI) to standardize the interaction between a PC's firmware and its operating system. UEFI also standardizes the operating system's loading method and the execution of pre-boot applications, which promises to reduce the software conflicts that undermine system stability. Firmware is an oft-neglected sector of the PC industry that typically does not get the same press that rollouts of new processors and operating systems attract. UEFI will shoulder much of the load that had been carried by the existing BIOS software in the first major revision of the method for writing firmware. A UEFI 2.0 specification is due out in the near future, and will likely pervade the PC landscape by 2007. The United EFI Forum, a group that draws support from Intel, AMD, and Microsoft, produced the new specification, building on Intel's Extensible Firmware Interface specification 1.1. The transition is expected to gather steam this summer, when Microsoft releases Windows Vista, which is expected to be compatible with firmware based in both BIOS and EFI. UEFI 2.0 is not likely to catch on in earnest until 2007, however, given that it is only rumored to boot 64-bit EFI operating systems. Once it is standardized, Dell is expected to convert to UEFI, said the company's Dick Holmberg. UEFI also allows incremental changes to be made bit by bit, while the rest of the package remains unchanged, minimizing PC instability. The new interfaces also have such appealing features as a boot manager, which enables PCs to boot from multiple devices and toggle between operating systems, as well as a network stack that links a PC to a network before its operating system has loaded.
Click Here to View Full Article
to the top

DOE's Office of Science Awards 18 Million Hours of Supercomputing Time to 15 Teams for Large-Scale Scientific Computing
U.S. Department of Energy (02/01/06)

The Energy Department has granted researchers 18.2 million hours of computing time at some of the fastest supercomputers in the world to advance a host of research projects through its Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program. "Through the INCITE program, the department's scientific computing resources will continue to allow researchers to make discoveries that might otherwise not be possible," said Energy Secretary Samuel Bodman. INCITE has already funded research in fields such as chemistry, astrophysics, and genetics, and will support new initiatives in disease research, aerospace, molecular simulations, and atomic-level protein structuring throughout the coming year. INCITE is also encouraging proposals from the private sector for the first time in its three years. Each of the last two years, the Energy Department has only issued three grants, though it is offering 15 this year thanks to an expansion of its program to include five supercomputers in the Argonne National Laboratory, Lawrence Berkeley National Laboratory, Oak Ridge National Laboratory, and Pacific Northwest National Laboratory. The department's call for proposals received responses from 11 scientific disciplines requesting more than 95 million hours of computing time. Boeing, DreamWorks Animation, General Atomics, and Pratt Whitney were the four grant recipients from industry, with the remaining 11 grants awarded to colleges and universities. Of the grant recipients already using their allocation, a University of Chicago group is simulating cosmos accretion, another is studying numerical simulations of flame, and a team from the University of Washington is cataloging the dynamical shapes of proteins using an IBM supercomputer.
Click Here to View Full Article
to the top

Self-Improving Software
Technology Research News (01/30/06)

Software that is able to learn on its own, and without a separate training process, could result from new algorithms developed by researchers at Princeton University. The algorithms are able to learn from data that they do not know anything about beforehand, and then make adjustments to handle such data. The researchers designed the algorithms to focus on the way pieces of data represent a range of possibilities, instead of the details of the data. Algorithms are able to learn from individual pieces of data, and make an improvement after dealing with a few samples. The algorithms have the potential to allow software to alter its default configurations on its own while it learns how it is used. The research, "Self Improving Algorithms," was presented at the ACM-SIAM Symposium on Discrete Algorithms, January 22-24, 2006.
Click Here to View Full Article
to the top

Words Help Us See and Talk
University of Chicago, Ill. (01/31/06)

Research in January's Proceedings of the National Academy of Sciences suggests that language may determine half of what we visualize. The study wades into the controversial issue of whether language influences perception, but adds to the discussion by suggesting that what we speak affects the right half of the visual field. The left hemisphere of the brain is largely responsible for processing language. "So it would make sense for the language processes of the left hemisphere to influence perception more in the right half of the visual field than in the left half," says Terry Regier, associate professor of psychology at the University of Chicago. Regier authored the paper, "Whorf Hypothesis is Supported in the Right Visual Field but not in the Left," along with Paul Kay, professor emeritus of linguistics and a senior research scientist at the International Computer Science Institute in Berkeley, Calif.; Richard Ivry, professor of psychology and director of the University of California at Berkeley's Institute of Cognitive and Brain Sciences; and Aubrey Gilbert, a graduate student in the Helen Wills Neuroscience Institute at UC Berkeley. The study was based on experiments on Berkeley undergraduates who were shown a ring of colored squares, which were the same color except for an "odd man out." Participants were charged with identifying which side of a circle (right or left) the odd-man-out was on, as the odd-man-out was given either the same name as the other squares (e.g. a shade of "green" while the others were a different shade of "green") or a different name (e.g. a shade of "blue" while the others were a shade of "green"). Responses came more quickly when the odd-man-out had a different name and was in the right half of the visual field.
Click Here to View Full Article
to the top

Ontology Construction From Online Ontologies
University of Southampton (ECS) (01/30/06) Alani, Harith

Harith Alani of the University of Southampton's Electronics and Computer Science Department posits that the high cost of building ontologies is a major obstacle toward their wide acceptance, and notes that existing ontology reuse, though less expensive, is a challenging proposition because of the relative newness of ontology reuse tools. Alani notes that the emergence of ontologies and libraries for storing and indexing ontologies on the Web, along with search engines that can expedite the search and retrieval of online ontologies, is a positive development, and he proposes a system for automatically building ontologies through the identification, ranking, and integration of online ontological fragments. The first step toward constructing an ontology is to identify terms to be represented in the ontology, and Alani's proposed system would analyze the ontologies retrieved in a search query to obtain as much representational data as possible about the given term. If an excessive number of ontologies is found, the system would rank them by certain criteria, such as the degree to which they represent concepts in the search query, user ratings of the ontologies, or how well they address the needs of specific assessment tests. The system would either take the ontology as a whole or only the segment describing the term in question, and then compare the various ontologies it finds to identify any additional representations that can be blended into the first ontology, resulting in an ontology that is richer and more refined than any existing ontologies. Evaluating the resulting ontology may be necessary to ensure it fulfills some minimum quality requirements. "Facilitating reuse of other people's ontologies should encourage more individuals and organizations to participate in the Semantic Web," Alani reasons.
Click Here to View Full Article
to the top

Forum Tackles Internet Regulation
ISN--Security Watch (01/31/06) Lyman, Eric J.

Representatives from leading high-tech companies around the globe met at Italy's Ministry of Culture in Rome in late January to discuss online security, reaching a general consensus that the less the government gets involved, the better. The International Conference on the Future of the Digital Economy was hosted by the Organization for Economic Cooperation and Development. Attendees noted that companies have taken security-related steps independently to avoid government intervention. "If companies take steps to assure security and legal issues, then there will be no need for government regulators," said David Sifry, the president of Technorati, a company that monitors traffic on Internet blog sites, on the sidelines of the event. Representatives from Google were on hand to discuss their much-maligned Google Book Search initiative, noting that only books in the public domain will be fully available. "We are operating on the assumption that regular copyright law is the regulatory structure that is relevant here," said Patricia Moll, Google's European policy manager. "We are not breaking any copyright laws and we don't see there being a need for additional regulation in this area." Jung Ju Kim, the CEO of South Korean multiplayer game developer Nexon Corporation, weighed in, saying, "We believe the government should step in only when there is no other way." But speaking at the event, Italian Innovation and Technologies Minister Lucio Stanca said that the government will indeed step in when needed. "The government's job is to protect its citizens in every way possible," Stanca said. "If the need arises in the future, that could include increased regulation of the online world." Italy in the past has shown a willingness to get involved in Internet affairs, strictly limiting the use of its .it domain, policing defamatory Web sites hosted in Italy, and limiting the sale of illegal products, as well as goods linked to extreme political groups over the Internet.
Click Here to View Full Article
to the top

The Race Is on to Debug Dual-Core Deadlocks
Software Development Times (02/01/06)No. 143, P. 1; DeJong, Jennifer

Deadlocks or race conditions that are hard to detect can cause a multithreaded application to halt in its tracks, but dual-core desktop processors from Intel and AMD are renewing interest in developing an antidote to such errors. Multithreaded applications contain threads that frequently vie for shared resources, along with variables and functions; in one example, a deadlock can crop up when a multithreaded application locks a resource and does not unlock it once its task is finished, causing a system freeze, according to Coverity's Andy Yang. Competition between threads for resources when the application fails to specify the order in which the threads can access the resources gives rise to race conditions, and Intel engineer James Reinders says synchronization can address this flaw. Multithreaded applications perform tasks concurrently, and avoiding errors requires the developer to know a program's resource-sharing principles. "Don't write multithreaded code if you don't know what you are doing," recommends Fortify CTO Roger Thornton. Collisions can be reduced through the use of "thread-safe" tools such as C# and Java, while Ken Cowan of Compuware says concurrency issues should be tackled in the design stage. Thornton believes multithreading should be used to increase performance, not just for its own sake, and adds that "you have the responsibility to weigh the trade-offs." Even when concurrency is necessary, developers need to consider a way to distribute the work among multiple processors without generating excessive synchronization, and tasks that do not exhibit a high degree of interdependence are the best candidates for parallelization.
Click Here to View Full Article
to the top

Robot Special: Walk This Way
New Scientist (02/04/06) Vol. 189, No. 2537, P. 40; Ball, Philip

Only recently has the concept of passive walking robots--machines that can imitate a human's entire gait cycle by relying on leg motion and gravity rather than motors--gained prominence, and an important sign of the technology's increasing distinction was the unveiling of three passive walkers at an American Association for the Advancement of Science conference last February. A team led by Cornell University's Andy Ruina built a robot that mimics people's ability to pump their gait by pushing off their back foot at the beginning of each step by using a spring in each lower leg; a small motor stretches the spring, which causes the ankle joint to flex when released. Another robot, built by MIT researchers supervised by Russ Tedrake, can sense the tilt of its body and other factors in order to "learn" how to walk, using an on-board computer that adjusts command signals transmitted to electric motors that flex the ankles. The third robot was the product of the Delft University of Technology in the Netherlands: The device is driven by compressed-air actuators in the hips, and features an ankle design influenced by skateboard suspension principles that adds stability. All three machines represent a significant advance because they simplify leg design and control, and lower energy consumption. Passive walkers' inability to steer or avoid obstacles, as well as their difficulty in dealing with uneven or pliable surfaces, makes them inappropriate models for robust walking robots, according to Carnegie Mellon University roboticist Chris Atkeson. He believes the best solution will mix principles of both passive dynamic walkers and powered walkers, and expects practical robots will be driven by most if not all of their joints. Passive-walking robots are also inspiring research into advanced prosthetics that could perhaps reduce the energy cost to the wearer and amplify human performance.
Click Here to View Full Article
to the top

Unified--and Agile
Software Development (01/06) Vol. 14, No. 1, P. 49; Ambler, Scott W.

Ambysoft software process improvement consultant Scott Ambler believes software development could be revolutionized by combining agility with the Unified Process (UP), and his work in this area has yielded the Agile Unified Process (AUP), a simplified approach to software development based on the Rational Unified Process (RUP). AUP shares with RUP a serial, iterative nature and an architecture-centric scheme, but the chief difference is AUP's simplicity. AUP possesses a small number of deliverables--source code, a regression test suite, and other artifacts that must be produced as part of the system--which can speed up time to market. AUP's model discipline consolidates RUP's Business Modeling, Requirements, and Analysis and Design disciplines, and less up-front modeling is required in AUP. In AUP, RUP's Configuration and Change Management discipline is supplanted by a Configuration Management discipline, while RUP's change management activities are incorporated into Model and Project Management disciplines. Ambler notes that agile techniques such as test-first programming, code and database refactoring, continuous builds, and continuous regression testing are adopted by AUP's Implementation discipline. The need for close collaboration between developers and stakeholders and agile modelers is explicitly depicted, as is database and administration initiatives. AUP encompasses an evolution of the Project Manager's role to reflect agile project management strategies.
Click Here to View Full Article
to the top

To submit feedback about ACM TechNews, contact: [email protected]

To unsubscribe from the ACM TechNews Early Alert Service: Please send a separate email to [email protected] with the line

signoff technews

in the body of your message.

Please note that replying directly to this message does not automatically unsubscribe you from the TechNews list.

ACM may have a different email address on file for you, so if you're unable to "unsubscribe" yourself, please direct your request to: technews-request@ acm.org

We will remove your name from the TechNews list on your behalf.

For help with technical problems, including problems with leaving the list, please write to: [email protected]

to the top

© 2006 Information, Inc.