The American Association of University Women estimates that about 20 percent of IT professionals and fewer than 28 percent of computer science graduates are women, who also constitute a mere 19 percent of engineering graduates, according to the National Academy of Engineering. American educators' relatively low expectations of females' performance in math, as well as U.S. students' lower math performance compared to foreign counterparts, are partly to blame for this situation. Ptak, Noel & Associates IT analyst Jasmine Noel also points to girls' general perception of an IT career as a "geeky" pursuit characterized by social isolation, even though IT is starting to break out of this stereotypical mold and incorporate more business/communication skills. "That mix of problem solving, technical knowledge and interpersonal interaction needs more emphasis if more women are to be attracted into IT," Noel says. A career awareness survey from The Futures Channel a few years ago determined that young girls considered teaching, secretarial, and nursing to be their top three career possibilities, while a more recent Lemelson-MIT Invention Index finds girls' preferred career choices to be, in descending order, acting (32 percent), music (24 percent), athletics (22 percent), president (17 percent) and science/innovation (a paltry 10 percent). Increasing women's participation in IT, the sciences, and engineering is the goal of numerous organizations such as the Anita Borg Institute and the Math/Science Network. "We need to spread the word to young girls about the importance of taking all the math and science they can in high school in order to maximize their career choices," explains Math/Science Network President Teri Perl. Noel says women must be taught that technology can be a vehicle for business career advancement rather than as something programmers need to know.
Click Here to View Full Article

U.S. capabilities in decryption and other narrow fields are threatened by the fast rise in clustered supercomputing, warned the National Research Council at ACM's SC2004 supercomputing conference. Clustered systems now make up 296 of the 500 fastest supercomputers, but are not as good as more expensive custom-built machines in handling some mathematical operations, while University of California professor Susan Graham noted that national expertise in other technologies such as vector supercomputing could atrophy as more money and development go toward clustered supercomputers. The group suggested federal agencies and Congress should fund more varied supercomputing research and cooperate on defining their specific needs. Currently, individual agencies think about supercomputing in terms of being a customer, but do not consider their responsibility in shaping the future supply of supercomputing technologies, Graham said; with dialogue between agencies, the government will be able to ensure some program is in place to research necessary technology. IBM's Dave Turek said the marketplace was still the best overall approach to discovering and promoting technology, and warned against siphoning government funds from successful technologies to uncompetitive ones. IBM is currently working to link conventional processors so that they emulate a vector supercomputer, such as the massive NEC-built Earth Simulator. The National Research Council report warned that vector computing was in danger in the marketplace because the applications were increasingly narrow. Penguin Computing CTO Don Becker, whose firm was a pioneer of Beowulf Linux clusters, said the clustered approach was rapidly becoming more advanced and that only a handful of problems would be out of reach of cluster supercomputers in five years' time.
Click Here to View Full Article

The Global Information Grid (GIG) is conceived by the Pentagon as a separate, secure Internet that blends military and intelligence activities so that soldiers anywhere in the world can instantly receive information about foreign and unfriendly troop movements. Supporters say the GIG is key to effecting the U.S. Armed Forces' transition to net-centric warfare, a transition that is needed in order to combat increasingly faceless, decentralized enemies. Billions have already been poured into the GIG's development, and Pentagon documents suggest that the network's software and hardware costs over the next 10 years or so could top $200 billion: Supplying the GIG's connections is expected to cost at least $24 billion over the next five years; data encryption is estimated to cost $5 billion; the radios needed for the net will probably cost $25 billion; and tens of billions more will be spent on the required satellite systems. The Pentagon told Congress five years ago that the GIG project will likely take 25 years or more to be realized. The biggest obstacle the GIG initiative could face is cultural, as all four military services have long-standing traditions of building their own technology and jealously guarding it from each other, which would be rendered obsolete by the unified force the war net seeks to facilitate. The project will also need to address a major technical challenge--providing adequate bandwidth, which must exceed the amount of bandwidth employed at the height of the Iraq campaign last year by a factor of 40 or 50, according to estimates from the Rand Corporation. However, the Congressional Research Service notes that only one-tenth of the necessary bandwidth may be available to the Army, citing a "lessons learned" report from Iraq concluding that "there will probably never be enough resources to establish a complete and functioning network of communications, sensors, and systems everywhere in the world."
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.)

Sun Microsystems' proximity communications technology won the overall Gold prize in the Wall Street Journal's 2004 Technology Innovation Awards: Sun researcher Robert Drost began working on better chip-to-chip communications as a Ph.D. student at Stanford University, but discovered the traditional wired approach had already been poured over by colleagues. The wireless approach Drost and co-researchers are pursuing could potentially boost computing operations 10-fold or more, is a cornerstone of Sun's Defense Advanced Research Projects Agency supercomputing work, and has been detailed recently in preliminary findings. The proximity communications technology still has several critical obstacles, including the precise alignment and signal transmission required between two chips, not to mention configurations where up to six chips are connected. In the network security field, judges chose Sana Security's Adaptive Profiling Technology for its ability to recognize anomalous network activity as a precursor to malicious software execution. Flarion Technologies won the Gold award in the network and Internet technologies category with its Internet protocol-based mobile broadband network, which some of the experts saw as an alternative to 3G networks. IBM's Marvel multimedia search technology won an award for the way in which it combines text, concept model, and metadata search techniques; such an approach would dramatically reduce the burden of labeling, indexing, and searching for multimedia files. The wireless category Gold award went to MeshNetworks for its ad hoc command system that provides both communication and location capabilities to emergency responders deployed in disaster areas. Witten Technologies won in the software category for a radar tomography application that rapidly analyzes underground radar images for engineering, construction, and environmental purposes.

A team of researchers has created a method of quantifying the aesthetic properties of documents so that they can be automatically converted to different formats without deviating from the original intent. Xerox researcher Steven Harrington and collaborators from the Rochester Institute of Technology presented their work at the recent ACM Symposium on Document Engineering. The researchers say designers can automate the layout of an advertisement for different types of media, such as on a printed piece of paper, a Web page, or mobile phone screen. Each media offers different opportunities to express the original intent, either through bold headlines, Web animation, or in the case of mobile phones, through sound and speech. The team identified over 150 measurable values that can be used to express the document's intent, including typeface and font, density of text, and image color and position. Each of these values can be categorized into a set of intents, so that a transformation matrix can automatically guide a document from one media to another. Layout proved to be a particularly complex and important aspect of document intent, and the team isolated several factors that resulted in good layout, including alignment, regularity, balance, and positioning white space. Harrington warned that just one bad aesthetic feature could ruin the entire document's appearance, but that the system and set of influencing factors the researchers collected provided a fairly good basis for automatic document conversion with consistent intent.
Click Here to View Full Article

Penn State researchers have created a new technology that filters out queries for unauthorized information instead of the data itself, thereby eliminating the need for individually built access control modules in databases. Because these databases no longer have to process illegitimate requests, response is faster, says Penn State assistant professor Dongwon Lee, who presented the research with colleagues at the ACM Conference on Information and Knowledge Management. Also, QFilter can be used with off-the-shelf databases without making significant changes to those products since queries are filtered out before they reach the database. Other methods of preventing access to unauthorized information include view-based technology, where separate views are created for each user. Though that allows fast response times, the view-based approach requires an inordinate amount of maintenance and storage as the user base expands. QFilter, on the other hand, provides fast response along with optimal storage and maintenance requirements, says Lee. QFilter makes use of non-deterministic finite automata computation, which keeps access control policies in centralized areas and blocks particular parts of a query that ask for unauthorized information. The researchers say QFilter will be further refined and have accompanying applications.
Click Here to View Full Article

Academic and legal experts warn that proposed revisions to the Copyright Act of 1998 that the standing committee on Canadian Heritage resubmitted to the World Intellectual Property Organization (WIPO) last week would exact a heavy, unfair toll on consumers. The revisions imply that all content on the Internet constitutes intellectual property, and cannot be printed out or copied by users without an Access Copyright fee, unless there is an explicit legal notice stating that the material may be used. "It means that the bulk of sites used in educational settings--resources designed by museums, libraries, universities, experts of various kinds--that are intended for educational uses may be levied," says Queen's University English professor Laura Murray, who condemns such a measure as "perverse." The committee also recommends that copyright holders be granted the right to order an ISP to remove online content on demand, without any due process for defense. University of Ottawa professor Michael Geist and Canadian copyright lawyer Howard Knopf blame this state of affairs on the committee's refusal to acknowledge evidence that opposes the demands of the Access Copyright branch of the book publishing industry and the Canadian Recording Industry Association. Knopf contends that his failure to translate material favoring a more balanced strategy into French led to its rejection by the committee. He notes that the committee ignored clear proof that a levy on blank CDs as a means to reimburse artists for pirated material would increase 100 percent. "The danger of WIPO is that it...replaces social rights with absolute rights," Geist argues. The WIPO treaty has also been under fire in Europe, while the U.S.'s Digital Millennium Copyright Act has been challenged in court for violating free speech rights and hurting security research and development.
Click Here to View Full Article

Projects to make computers more accessible to Africans who do not know English, French, or the other major languages programmed into desktop systems promise to open up new markets for IT vendors, close the gap between Africa's technology haves and have-nots, and preserve hundreds of increasingly obscure dialects from extinction. Though Internet cafes are establishing a presence in even the smallest African communities, the language barrier limits their use to only the wealthy educated classes, a situation that experts say must be rectified by developing software and adapting keyboards to speakers of Swahili, Amharic, Yoruba, Sesotho, Hausa, and other African languages. African Languages Technology Initiative executive director and linguist Tunde Adegbola has developed a special keyboard that omits letters that do not appear in the Yoruba alphabet, increases the prominence of vowels, and adds accent marks and other symbols. Adegbola is releasing the keyboard free of charge to prominent Yoruba speakers in the hope that they could bring in venture capital. His latest project is to devise voice recognition software that can translate spoken Yoruba into text. Meanwhile, researchers at Ethiopia's Addis Ababa University have created a text-messaging system for Amharic speakers that could be employed as a development tool for farmers in far-flung regions. A system that translates English into the South African dialects of Afrikaans and Southern Sotho is under development by researchers at the University of the Free State's Unit for Language Facilitation and Empowerment. And Microsoft is working to embed Swahili into its popular programs, partly as a way to stem the spread of open-source operating systems such as Linux, which are growing in popularity.
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.)

New Jersey Institute of Technology (NJIT) professor Michael Bieber is leading an effort to deliver electronic material regardless of where a query is being carried out. The method involves locating articles containing clickable author- or keyword-associated "anchors" or icons that link to content in the book catalog or within other databases, videos in digital libraries, notes added to the articles by a local professor, the author's Web page, or even MapQuest or similar services. Bieber explains that project participants are employing a system-specific "wrapper" that scans the display screens such as where the title and author are located, and analyzes the screens lexically, assigning anchors to important phrases within registered glossaries; clicking on the anchor causes the link to transmit a request to an information provider, and generates an ad-hoc link list as identified by a relationship rule or glossary entry. Bieber notes that bibliographic data at the end of articles is accessed via link resolvers, although sources that do not have metadata or are not written in XML will be harder to access. Because the project is open source, wrappers produced for data from one vendor or particular data set can be employed by other libraries with few alterations. The NJIT project also involves the development of a search engine that compiles searches from multiple engines and generates clusters with key phrases, and a collaborative filtering element that employs user data to supply customized information. The initiative's proof of concept stage involved the participation of the National Science Foundation's Digital Library and NASA, and project participants have won authorization from the Association for Computing Machinery and EBSCO Gale ProQuest to use their data for research. The NSF has contributed to the approximately $2 million in funding the NJIT project has received.
Click Here to View Full Article

India's Center for Development of Advance Computing (C-DAC), the Indian Institute of Information Technology, the Indian Institute of Technology, and Mysore are collaborating on speech-to-speech translation technology that will enable mobile phone users who speak different languages to communicate. The technology marries two translation processes, speech-to-text and text-to-speech: The speech-to-text process takes place when one caller's message is converted into audio signals that are digitized and analyzed to glean key features of the spoken word, which is then deciphered through analysis by artificial intelligence methods trained to understand both dialects' linguistic structure along with the syntax, semantics, and pragmatic knowledge; the text-to-speech process transforms the translated text into audio signals through the extraction of knowledge from databases, computer servers, and Web pages. Speech and language communication researcher Shyam S. Agrawal says the software currently suffers from a one- to two-second latency between speech at one end and translation at the other, but he says this latency will disappear once the conversation goes forward. He also predicts that future mobile phones will have embedded processors that eliminate the need for computers and servers in speech-to-speech translation. Agrawal foresees the technology fueling consumer demand for home-based voice synthesis or voice recognition, and being used in such applications as voice-controlled appliances and enablement technology for physically handicapped people. It is projected that systems will help access data in a language known anywhere on Earth by the end of the decade with the advent of cross language information retrieval and speech-enabled online information retrieval.
Click Here to View Full Article

United Nations Secretary-General Kofi Annan has announced the appointment of about 40 government, private sector, and civil society representatives from all regions of the world to the United Nations Working Group on Internet Governance. The establishment of the Working Group comes after the first phase of the World Summit on the Information Society last December, and the group will be tasked with addressing a number of issues related to Internet governance ahead of the second phase of the summit, to be held next year in Tunis. Under the leadership of Chairman Nitin Desai, special adviser to the secretary-general for the World Summit, the working group will meet for the first time in Geneva Nov. 23 to 25. The working group is charged with defining Internet governance and addressing relevant public policy issues, as well as reaching an agreement on the roles of governments, international groups, the private sector, and civil society in the area of Internet governance. "We come to this process as facilitators, and will strive to establish a dialogue of good faith among all participants," Desai says.
Click Here to View Full Article

The new Mydoom worm variant takes advantage of a security vulnerability that does not yet have a patch. Many programmers hunt for bugs in software, with most of them trying to help users by disclosing the flaws so they can be fixed. But the disclosures can help malicious hackers as well and cyberethics experts say the issue needs to be addressed. The new Mydoom variant arrives in an email message that seems to come from PayPal, and clicking on the link within releases the worm into the computer. The worm then sends out more email and leaves a back door open on the computer so that attackers can install other programs. Hackers exploited this flaw in Microsoft's Internet Explorer browser very quickly since it was only discovered in late October. Notification was posted on mailing lists and some people promoted it to try to get Microsoft to fix it more quickly. Product teams started working on a patch at once but were not finished by the time the worm began attacking. Microsoft has called for more use of the "responsible disclosure process," which involves private notifications to vendors of potential flaws until the patch is released. However, some researchers argue that this gives vendors an excuse to put off patches, which could place users at risk; they say full disclosure gives users a better chance to counter virus threats while also putting pressure on vendors to release software that is more secure. Others, such as Black Hat President Jeff Moss, says a middle ground is best, in which vendors are given a certain amount of time to fix problem before the public is notified. He says, "I want the vulnerability fixed and then I want all the security professionals notified."

Research funded by the National Science Foundation has led to the development of the Hug, a robotic pillow designed by Carnegie Mellon University researchers that employs sensing and wireless phone technology to provide elderly people with warmth and tactile sensations approximating the embrace of family members. The velour-covered device, which is configured to resemble a human torso with two arms reaching out, can be used by the sender (a grandchild, for instance) to send hugs to the receiver (a grandparent). The grandchild would squeeze one of the Hug's "paws" and utter the grandparent's name; the utterance would be picked up by microphone and fed to voice recognition software, which would identify the name and match it to a preset phone number consistent with the recipient's Hug. The grandchild's device would send a call to the grandparent's counterpart, which would emit a visual and audio signal that the grandparent could respond to by squeezing the Hug's left paw and saying hello. Then the grandchild would hug the device, and the sensations would be converted by sensors into a data stream and transmitted to the grandparent's device, triggering built-in motors to generate corresponding vibrations while thermal fibers radiate warmth. Research shows that older people are most often in need of emotional support, and the Hug was designed to provide such support remotely, according to CMU professor Jodi Forlizzi. CMU researcher Francine Gemperle notes that the Hug can have broader applications beyond the grandchild-grandparent relationship it was originally designed for. She acknowledges that the device "would need to go through product development, where people may want to change its appearance and make it more adaptable to different-sized people."
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.)

ICANN on Friday will issue a formal request for proposals from companies wishing to oversee the .net domain name after VeriSign's contract expires at the end of June 2005. "Thirty-seven of the top 100 Web sites rely on .net, and about 37 percent of all e-commerce relies on .net to get to its destination," says VeriSign vice president of government relations Tom Galvin, noting the importance of the domain's continued stability. Along with VeriSign, expected bidders include Germany's Denic, the registry for the .de domain, as well as .biz registry NeuLevel and Afilias, which manage more than 5 million .info and .org domains. ICANN and VeriSign have shared a tumultuous relationship of late. In October of last year, ICANN forced VeriSign to shutter its Site Finder service. In May, a federal judge dismissed a suit filed by VeriSign accusing ICANN of improperly trying to interfere with new offerings. A pending suit filed by VeriSign in California State Supreme Court accuses ICANN of violating the 2001 .com Registry Agreement by prohibiting a number of services, including Site Finder and Wait Listing Service. ICANN will use a third-party to manage bidding and selecting of the domain, and the final choice will be submitted to the U.S. Department of Commerce for approval. Galvin says his company's ongoing legal issues with ICANN should not be a factor in the discussions over .net. He says, "We hope the decision on .net will be based on merit."
Click Here to View Full Article

The Segway Human Transporter (HT) has yet to live up to its hype as a life-changing transportation technology, although there are signs that Segway is finally starting to apply the two-wheeled scooter's experiential advantages as a selling point. No new Segway products in three years has bred skepticism that the company has enough funding to support another project and sustain the HT at the same time, but Segway CTO Doug Field reports that Segway has been "innovating" intensely, generating hundreds of concepts each year to find the next potentially big product. The company recently unveiled the Centaur, a prototype for a four-wheeled, gyro-stabilized all-terrain vehicle that can change its shape to run on two wheels and move faster than the HT. The Centaur's sensor- and gyroscope-equipped base is taken directly from the HT, while the steering column is outfitted with two sensors. One sensor reads the throttle to calculate speed, while the other reads handlebar movements to determine the angle of turn. This data is routed to computers in the base, and combined with readings from the tilt sensors and gyros by the control boards, which keep the Centaur upright by adjusting the velocity of the independent rear wheel motors 100 times a second. Field supports his company's design and development of products whose wide market appeal is not guaranteed, arguing that "Real breakthroughs almost always come from technical exploration and people trying to solve problems before they even know why they might want to solve them." Barring disagreement among Segway's market analysts, the Centaur could make its commercial debut by late 2005.
Click Here to View Full Article

Johns Hopkins University professor Avi Rubin has become a leading figure in the debate over the security of electronic voting thanks to his analysis of Diebold AccuVoteTS voting system software, which uncovered fundamental vulnerabilities that have sparked demands for a way to verify electronically cast votes. Conspiracy theorists contend that vendors are intentionally rigging the voting systems in order to support their political agendas, but Rubin dismisses such assertions, arguing that it is more critical to investigate the potential for security problems. Experimentation demonstrates that malicious code can be concealed in other code with relative ease, while detecting such code is a much harder prospect. Observations that e-voting systems appear to be performing well irritate Rubin, who explains that such an assessment is only applicable to issues of functionality, not security. He makes a strong case for the incorporation of voter-verifiable paper trails, which will make it more difficult for voting machines to rig elections as well as make mistakes. The Johns Hopkins researcher says he and other academic professors have written a joint proposal to the National Science Foundation to research the design and construction of secure, accurate, verifiable, and accessible voting machines, a project that requires an estimated grant of $10 million. He firmly supports an open-source model in the design and development of e-voting software, arguing that "the appearance of a lack of transparency is not something we can tolerate in voting systems." Rubin notes that public awareness and discussion of e-voting security issues is spreading thanks to recently proposed federal laws that champion open-source, voter-verifiable paper, and predicts that within 15 years, "we are going to have systems that are provably accurate and universally verifiable, systems where code can be published on a Web site for which high-school students could write a program to verify that would prove the results were correct without revealing how any individual voted."

Hopes that networking technology would make the U.S. military a leaner, more effective fighting machine were undercut by its disappointing performance in the Iraq War's ground campaign, which was characterized by a lack of situational awareness among ground forces due to a "digital divide." Such conclusions are documented in a largely classified assessment of the war being readied by Rand. The deployment of cutting-edge technology on the battlefield has been heavily promoted by the Pentagon as a enabler of "force transformation," in which soldiers would be networked into smaller teams thanks to unmanned planes and ground vehicles using advanced sensing, targeting, imaging, and communications capabilities. The Iraq campaign was a highly streamlined version of this model, involving sensors mounted on aircraft and satellites to support front-line troops; but though the implementation went smoothly, communications breakdowns became endemic once the actual invasion started. Ground force units would often outrun the range of their microwave-based communications relays, while commanders coordinating the war in Kuwait and Qatar sometimes found themselves inundated by an overwhelming flood of sensor input. System breakdowns made it difficult for troops to download critical aerial reconnaissance photos sent by high-level commanders. Owen Cote of MIT's Security Studies Program says problems such as these indicate that "If there is this 'revolution in military affairs,' and if this revolution is based on technologies that allow you to network sensors and process information more quickly and spread it out quickly in digestible form, we are still just scratching the surface of it." Certain force transformation proponents blame networking technology's shortcomings in Iraq on an outdated, vertical command and control systems model, when a horizontal model would have been more effective.
Click Here to View Full Article
(Access to this site is free; however, first-time visitors must register.)