ACM TechNews (HTML) Read the TechNews Online at: http://technews.acm.org
ACM TechNews
June 25, 2007

Learn about ACM's 2,200 online courses and 1,100 online books
MemberNet
CareerNews
Unsubscribe

Welcome to the June 25, 2007 edition of ACM TechNews, providing timely information for IT professionals three times a week.


HEADLINES AT A GLANCE:

 

Planned Worker ID Called Vulnerable
San Francisco Chronicle (06/25/07) P. A4; Lochhead, Carolyn

Various proposals to control illegal immigration rely on an electronic employer verification system that a Department of Homeland Security study criticizes as susceptible to identity theft, employer abuse, data inaccuracies, and privacy breaches, which could only be addressed through heavy enforcement. The detection of identity theft is not designed into the Web Basic Pilot system, and the size of such a system seriously complicates practical application, according to experts such as [[SRI International]] Computer Science Laboratory scientist Peter Neumann, who cited the approach's many shortcomings in recent testimony before Congress on behalf of ACM. He said lawmakers often have unrealistic hopes for technological solutions to social problems, and referred to a series of government software development blunders that include "many highly visible projects that have been late, over budget, or indeed abandoned after many years and large expenditures." Others mentioned that it was possible to build such a system, acknowledging that it would likely cost billions of dollars and require an immense technical effort. The DHS study concluded that the system is vulnerable to anyone disguised as an employer to gain access, and Neumann said there is no doubt that criminals would start creating "phishing" emails claiming to be from the DHS requesting worker data from unwitting employers. Mike Aitken with the Society for Human Resource Management predicted that the increasing security of immigration documents will raise the likelihood "that U.S. citizens' identities are going to be stolen and fraudulently used for employment by those who don't want to come out of the shadows," to the degree that the situation "will be worse than what we have now." ACLU legislative counsel Tim Sparapani warned that the system would empower the government to refuse people the right to work on an unprecedented scale, while being ultimately ineffective. [[For more on Peter Neumann's testimorny, visit http://www.acm.org/usacm]].
Click Here to View Full Article
to the top


High-Tech Titans Strike Out on Immigration Bill
New York Times (06/25/07) P. A1; Pear, Robert

Representatives from leading technology companies have been pressuring lawmakers to increase the number of foreign workers allowed into the United States under the H-1B visa program, but have had limited success so far. Currently, an immigration bill in the Senate would expand the number of work professionals, but high-tech companies say the expansion is not enough. Those opposed to expanding the visa program argue that H-1B workers are allowing large technology companies to hire immigrant workers because they can pay them less than a similarly skilled American worker. Originally, the bill created a point system that would reward applicants based on their degree level and job skills, but tech companies opposed the system because it would prevent them from sponsoring specific applicants for specific positions. An upcoming amendment would set aside 20,000 green cards for immigrants with extraordinary abilities, such as outstanding professors and researchers or managers and executives of multinational organizations. The amendment would also give employers five years to adjust their hiring practices to the point system. The number of green cards for employer-sponsored immigrants would gradually decline from 115,000 during the first two years to 44,000 in the fifth year, with no employer-sponsored green cards after that. The Senate bill would, however, raise the limit of H-1B visas from the current 65,000 to 115,000 in 2008 with the possibility of a 180,000 limit in later years if labor demands require another increase. The number of foreign students educated in the United States and allowed to stay after graduation would also be raised. The amendment would double the number of exempt graduate degree recipients, as well as add 20,000 H-1B visas for people who hold an advanced degree from a university outside the United States.
Click Here to View Full Article
to the top


K-12 Alliance Launched to Reverse Declining Participation of Girls in Computing Careers
Business Wire (06/25/07)

The National Center for Women & Information Technology (NCWIT) announced the creation of a new coalition intended to boost the number of girls interested in pursing careers in computing fields. The NCWIT K-12 Alliance, composed of 19 organizations including ACM, will work to improve the visibility of girls' involvement in computing and information technology, remove obstacles preventing female participation in the field, improve computer education at the K-12 level, and raise awareness that strong computer skills create success in many other careers. The U.S. Department of Labor says that only 26 percent of IT workers in the U.S. are women and predicts that more than 1 million computing jobs will be added to the workforce by 2014. Surveys by the Higher Education Research Institute show an 80 percent decline between 1996 and 2005 in the number of incoming undergraduate women interested in computer science. "In the next seven years, women will account for more than half of the nation's workforce," says NCWIT CEO and co-founder Lucy Sanders. "If U.S. companies wish to maintain their competitive advantage in IT-related fields, they cannot afford to miss out on the input of half the population. Women can, and must, play a more significant role in building an innovative and technically trained workforce." The K-12 Alliance's first project will be to release a resource kit called "Gotta Have IT" that will contain posters, career information, digital media, and more for teachers to use in the classroom. The K-12 Alliance is also creating a permanent networking system to help K-12 Alliance members distribute information to educators, parents, and other K-12 influencers.
Click Here to View Full Article
to the top


Linux Coders Tackle Power Efficiency
CNet (06/25/07) Shankland, Stephen

Linux programmers are making Linux computers more power efficient by creating a "tickless" kernel that abandons traditional time-keeping methods for a more power-friendly technique. Intel is also developing a power saver, called PowerTop, that makes it easier for users to know what software is needlessly keeping a computer's processor running. "It makes a lot of sense," says Illuminata analyst Gordon Haff. "Raw, flat-out horsepower is less and less what the game's about--especially on laptops, which are becoming more common." The tickless kernel is already making its way into the Linux mainstream. Intel programmer and longtime kernel developer Arjan van de Ven says a laptop running Linux now consumes 15 to 25 percent less power when idling than it did three months ago. Processors consume a significant amount of power, frequently more than a 100-watt light bulb, and heat reduction techniques such as air conditioning in data centers and fans consume even more power. For some time now, processors have been able to enter power-saving modes when a user sets the computer to standby mode, but more can be done. Gigahertz-frequency processor cycles last less than a billionth of a second, which would allow a processor to enter and leave low-power states multiple times between two keystrokes of a fast typist, and software often needlessly keeps kernels active. The tickless kernel, instead of frequently checking for work to be done, schedules the hardware to interrupt when it knows a future job will require attention. The tickless kernel also allows for better use of virtualization, which allows multiple operating systems to simultaneously run on the same computer, which can reduce the total number of machines operating to save even more power.
Click Here to View Full Article
to the top


Hopkins Center to Help War Effort
Baltimore Sun (06/25/07) P. 1A; Emery, Chris; Connolly, Allison; Gorman, Siobhan

Johns Hopkins University will receive at least $48 million from the Department of Defense to develop computer systems that will help military and spy agencies process the massive amounts of intelligence information they collect. The grant is for new research focused on improving technology that can automatically translate and analyze speech and text from multiple languages, university officials announced. The technology would help overwhelmed intelligence analysts manage the huge amounts of information, often in Arabic, being gathered in Iraq and the war of terror. The Human Language Technology Center of Excellence is being established near Hopkins' Homewood campus and will be staffed by engineers, computer scientists, mathematicians, cognitive scientists, and linguists. Experts from the University of Maryland, College Park, and BBN Technologies will also participate in the research project. "It's really supposed to be a fresh look at this problem," says Gary W. Strong, the center's executive director. "This technology has hit a wall at this point." Mark M. Lowenthal, a former senior intelligence officer who oversaw language training across the intelligence agencies, says advances made in machine translation of conversations would be a major boost for intelligence agencies. "Everyone is busting their heads on machine translation," Lowenthal says. "That's sort of the Holy Grail." Carnegie Mellon's Alan Black, a researcher at the school's Language Technologies Institute, says current translation systems are slow and expensive and have difficulty translating words when not spoken clearly.
Click Here to View Full Article
to the top


New York Legislators Keep E-Voting Software in Public Hands
Computerworld (06/25/07) Songini, Marc L.

New York state voting activists are pleased that this year's New York senate and assembly session ended without changing the state's strict e-voting software escrow law. Activists were worried that pressure from the e-voting industry would force changes in the law that requires voting system vendors to place all source code and other related software in escrow for the New York State Board of Elections so it can be examined as necessary. The law also forces a voting system vendor to waive all intellectual property and trade secret rights if the software needs to be reviewed in court. Microsoft, whose Windows software is used in some e-voting devices, sought to amend the law to avoid the strict escrow provisions. New Yorkers for Verified Voting executive director Bo Lipari says concerned citizens created a swell of support in the legislature to ensure the law remained unaltered and that about 3,000 constituent calls created a forceful reminder to lawmakers of their commitment to strong voting laws. "The voting machine vendors have known for two years what our laws said," says New York state assemblywoman Barbara Lifton. "We're holding firm on our current state law which calls for open source code." Lipari says Microsoft's proposed changes would ruin the source code escrow and review procedures in the current law. Microsoft says it does not make its source code available for escrow under election law because of concerns that the code could be disclosed to third parties without adequate protections for intellectual property rights.
Click Here to View Full Article
to the top


When Computers Attack
New York Times (06/24/07) P. 4-1; Schwartz, John

Doomsayers have long been forecasting the digital equivalent of Pearl Harbor, when America's enemies attack U.S. computer networks in the hopes of crippling vital infrastructure, but experts claim the reality of a cyberwar scenario is considerably less extreme. Andrew MacPherson of the University of New Hampshire reports that, unlike physical attacks, recovery from cyberattacks requires less of an effort, given the resilience of the Web. Although the U.S. government has been preparing for a major digital assault, experts believe the United States gears up for cyberattacks every day through exposure to malware, meltdowns, glitches, and crashes, while there are very few points in the network where a single computer malfunction can cause a systemwide crash. Furthermore, human beings are also resilient, and a loss of services through one kind of medium can be offset through improvisation. Still, Danny McPherson with Arbor Networks thinks a full-scale cybertattack could have enormous ramifications, although he contends that the effects of cyberwarfare on the Internet will be much subtler than anticipated, in that "certain parts of the system won't work, or it will be that we can't trust information we're looking at." There is general consensus among experts that cyberwarfare is unlikely to resemble the recent blockage of online access to Estonian banks and government offices via distributed denial of service attacks, which was eventually attributed to tech-savvy activists protesting the relocation of Soviet-era war memorials, rather than the Russian government.
Click Here to View Full Article
to the top


A Sunny Hiring Season for Job Seekers
CNet (06/25/07) Olsson, Miriam

The job market is ripe for recent college graduates in the computer industry. The overall unemployment rate for the computer industry at the end of the last quarter was 2.1 percent, even lower than it was during the peak of the dot-com boom. The unemployment rate for software engineers is particularly low, down to 0.9 percent, according to the U.S. Labor Department, while the unemployment rate for recent technology grads is only 2 percent. In the battle to attract new computing graduates, companies such as Google, Microsoft, and IBM are spending significant amounts of time and money to recruit students. It is estimated that IBM spends more than $100 million on student activities annually. IBM has even gone so far as to create its own academic discipline, known as Services, Sciences, Management and Engineering (SSME), to ensure a flow of graduates with desired skills. Many companies are having trouble filling positions. Microsoft technical staffing manager Jeremy Brigg says "the pool of qualified folks in tech as a whole has shrunken in the U.S." Microsoft expects to high 2,500 students this year, both full-time and intern candidates, and would like to hire even more, but there are just not enough students with technical skills, Brigg says. Company recruiters often look for different skills. IBM, for example, looks for candidates with "preset skills," many of which can be found in SSME students, while Google prefers to hire early career candidates and shape them as the company sees fit. Despite the currently strong job market, there are some signs of challenging times to come. Challenger, Gray & Christmas CEO John A. Challenger cites the slowing economy, and Dell and Motorola both recently announced job cuts. However, Challenger notes that other technology sectors are still strong and growing.
Click Here to View Full Article
to the top


Human-Aided Computing
Technology Review (06/22/07) Greene, Kate

Microsoft researchers are trying to utilize some of the specialized, and often subconscious, computing power in the human brain to solve problems that have been difficult for computers to solve. Microsoft researcher Desney Tan and University of Washington graduate student Pradeep Shenoy have developed a system that uses electro-encephalograph (EEG) caps to monitor the brain activity of people looking at pictures of faces and objects. The researchers found that even when the subjects were not trying to distinguish faces from other pictures, there was still a subtle difference in brain activity. The researchers wrote software that examines EEG data and classifies faces and non-faces based on the subjects' responses. When one subject viewed an image once, the system was able to identify faces with up to 72.5 percent accuracy. The accuracy increased to 98 percent when eight people viewed a particular image twice. "Given that the brain is constantly processing external information," Tan says, "we can start to use the brain as a processor." Although the current research is mainly just proof of concept, eventually such face-recognition techniques could be used to search surveillance videos to quickly find images with faces and those without. Subconscious brain power could also be used to improve automated image searches by preclassifying objects to help a computer find the requested pictures more accurately.
Click Here to View Full Article
to the top


The Search for 'God's Number' in a Rubik's Cube
Boston Globe (06/25/07) Baker, Billy

Speedcubers are competing to solve a randomly scrambled Rubik's Cube the fastest, while computer science researchers at Northeastern University say the best-selling toy of all time can now be solved in 26 moves. Dan Kunkle, a doctorate student in computer science, and Gene Cooperman, a professor of computer science, used 128 CPUs running more than 63 hours to do most of the calculations for determining the minimum number of moves to solve the cube from more than 43 quintillion possible arrangements. The researchers spent more than a year and a half on the project at Northeastern's High Performance Computing Lab, and part of a $200,000 grant from the National Science Foundation on the research. "It has wide applications on mathematical group theory, enumeration, and search," says Kunkle. The research has practical applications such as scheduling for factories and air traffic control, he adds. Erich Kaltofen, a professor of mathematics and computer science at North Carolina State University who was not involved in the project, says the research is important because of the computing power and mathematical ingenuity that was used. "The problem is not finding the minimum number of moves to solve a Rubik's Cube, but to demonstrate that you can carry out such a gigantic combinatorial search," says Kaltofen. Kunkle and Cooperman will present their research in July at the International Symposium on Symbolic and Algebraic Computation in Waterloo, Ontario.
Click Here to View Full Article
to the top


Imaging System Brings Skeletons to Life in 3-D
Popular Science (06/20/07) Mika, Eric

Researchers at Brown University have developed a new imaging system for peering into living things at a desirable speed, resolution, and depth. The imaging system makes use of computed-tomography (CT) scanners to obtain detailed 3D views, and the fluoroscopy technique for turning a rapid succession of x-rays into video. Called CTX imaging, the new process can put 3D animations of bones in motion--walking, running, and jumping. After the traditional CT scan creates a 3D model of the bone structure of a subject, high-speed fluoroscopy records motion from two different angles, and then the two data sets are fed into image-processing software that combines them to generate animation in action from any angle. Resolution reaches a tenth of a millimeter and motion is captured at 1,000 frames per second. Researchers are using the room-size system to study how flight evolved in birds, and they believe it will be useful to orthopedic surgeons, who could look for better treatments for bone-, ligament- and joint-related injuries. The technology is unlikely to be incorporated into a pair of glasses in the near future, but a commercial version of the system for producing real-time video could hit the market by 2010, according to Elizabeth Brainerd, a biomechanics professor who heads the CTX program.
Click Here to View Full Article
to the top


Welcome to the World of Haptics for Industrial Applications
Basque Research (06/22/07) Perez, Rosa Iglesias

Vibrating cell phones, gaming controllers, and force-feedback control knobs in cars are just the beginning of haptic technology, which applies force, vibrations, and motion to connect users to computerized systems through the sense of touch. Interacting with or manipulating a screen is no longer limited to vision and sound, and with the emergence of PHANToM haptic and hand exoskeleton devices "you can feel or touch what you see." Current use of the devices can be seen in virtual modeling, medicine, education, assistive technology for the vision impaired, industrial design, and maintenance. In the industrial design field, haptic technology can be incorporated into computer-aided design (CAD) systems to allow designers to also feel forces and local stimuli similar to real situations during the assembly process. The Collaborative Haptic Assembly Simulator (CHAS) was developed to allow designers in different locations to grasp virtual parts and assemble them into a digital engine, for example. Operators at Labein in Spain and Queen's University Belfast in Northern Ireland used CHAS to collaborate in real time, enabling the operator in Belfast to grasp and feel collisions as the other operator assembled a part. Such collaboration over the Internet and interaction over distance could allow doctors to diagnose and operate on patients remotely, or an individual to shake another person's hand virtually.
Click Here to View Full Article
to the top


Ahead of the Tape
Economist (06/21/07)

Trading algorithms are now being developed to take news reports into consideration as they make their trading decisions. The computer programs that generate buy and sell orders can now process trades in thousandths of a second, and the new capability is viewed as a competitive advantage when lightning-quick decisions on the best electronic prices for trades need to be made. Morgan Stanley's Andrew Silverman says the news-perusing function has room to develop, and there are some concerns whether the software will understand the context of a headline. A third of all share trades in the United States involves algorithmic trading, which the consultancy Aite Group expects to account for more than half of all share volumes and a fifth of options trades in three years. And the research firm TowerGroup believes $480 million will be spent this year on developing technology for algorithmic trading. On Monday, the London Stock Exchange introduced an electronic system to accommodate the faster trades of algorithmic trading technology, and processing rose from 600 orders a second to 1,500. Britain's Financial Services Authority is also considering using the technology to search trading data for clues of suspicious trading activity.
Click Here to View Full Article
to the top


The New Metrics of Scholarly Authority
Chronicle of Higher Education (06/15/07) Vol. 53, No. 41, P. B6; Jensen, Michael

The digital movement of information assets is leading to a time where information is in great supply, and the way we establish authority, relevance, and scholarly significance is changing dramatically as a result, according to National Academies director of strategic Web communications Michael Jensen. It is critical that sense and structure be imposed on the massive flood of Web content on the horizon, and many technology-oriented thinkers believe artificial intelligence will fuel the "Web 3.0" trend. Jensen notes that algorithmic filtration models are starting to emerge, and cites the National Academies Press book-specific Search Builder and its Reference Finder as examples of Web projects that could provide insightful hints into the characteristics defining the trend. With the Search Builder, a researcher can choose terms from a chapter and generate term-pairs for submitting a search query to the National Academies Press (NAP), Google, and other services, yielding a precise answer. The Reference Finder is a prototype that allows researchers to paste in the text of a rough draft and retrieve related NAP books, based on algorithmically mined and ranked key terms. Jensen speculates that the Web 3.0 era will feature the emergence of intensely computed reputation-and-authority metrics, and probable elements of Authority 3.0 include publisher, peer reviewer, and commenter prestige; the portion of a document quoted in other documents; raw links to the document; valued links; nature of the language in comments; obvious attention; duration of a document's existence; types of tags assigned to the document, the terms used, and the taggers' and the tagging system's authority; percentage of phrases valued by a disciplinary community; context quality; relevance of author's other work; quality of the author's institutional affiliations; the reference network; and inclusion of the document in "best of" lists, syllabi, indexes, etc.
Click Here to View Full Article
to the top


Silicon Smackdown
Scientific American (06/07) Vol. 296, No. 6, P. 32; Frenkel, Karen A.

Creating a computer program that beats humans at the Asian board game "Go" is an enormous challenge, given the immense number of maneuvers open to players. The UCT (upper confidence bounds applied to trees) algorithm developed by Levente Kocsis of the Hungarian Academy of Sciences' Computer and Automation Research Institute and Csaba Szepevari of the University of Alberta in Edmonton can reportedly outdo the win rates of the best Go computer programs by 5 percent and compete with professional Go players on small boards, via an extension of the Monte Carlo strategy. The Monte Carlo approach assesses and ranks candidate Go moves by playing a large sample of random games, while the UCT algorithm concentrates on the maneuvers with the most potential. Kocsis explains that UCT must find a balance by testing alternatives that seem optimal at the time to find possible vulnerabilities and probing "less optimal-looking alternatives, to ensure that no good alternatives are missed because of early estimation errors." Kocsis projects that programs such as UCT could best professional human Go players within a decade. UCT has applications beyond games, because it can be applied to any problem that involves selecting the best option, provided the alternatives possess an internal tree-like organization and their values can be recursively computed. Targeted Web advertising, the optimization of channel allocation in cellular systems, and determining the best sites for industrial plants are just some of the potential applications for UCT.
Click Here to View Full Article - Web Link to Publication Homepage
to the top


Embedded Web Services: Making Sense of Diverse Sensors
Sensors (06/07) Culler, David E.; Tolle, Gilman

Web services promise to facilitate easy integration of diverse and distributed sensor networks, by employing the service-oriented architecture (SOA) model that is used to merge numerous data sources from physically scattered dynamic processes. The integration of diverse information sources transpires at three levels: The bottommost level is the communication medium that permits the exchange of data and control actions; the middle level is comprised of object and data representation; and the topmost level is service discovery, through which participants can squeeze value out of information exchanges by learning what actions other devices can carry out, as well as how specific actions can be facilitated. The majority of modern-day sensors are minuscule chips or circuit elements that are directly linked to powerful yet cheap microcontrollers with complex communication connections, and integrating these sensors into rich networks should be a simple prospect, because intelligence and communication are linked to each device. The integration of diverse information sources and processes is becoming commonplace, even in an online shopping site. This integration breakthrough was fueled by two developments. The first is a dramatic simplification at all three integration tiers, represented by the conversion of communication into the transference of sequences of characters between named endpoints on hosts, the streamlining of information representation to the recognition of nested XML-tagged sections, and reduction of the set of behaviors to GET and POST a sequence from, or to, a named endpoint. The second development is consistent SOA application in which applications are designed as a composition of services that is independent of protocol and deployment.
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

News Abstracts © 2007 Information, Inc.


© 2007 ACM, Inc. All rights reserved. ACM Privacy Policy.