The latest congressional wrangling over proposed anti-spam measures revolves around Sen. Charles E. Schumer's (D-N.Y.) bill to deploy a do-not-spam list modeled after the do-not-call registry the FTC started implementing in June. Jerry Cerasale of the Direct Marketing Association argues against such a measure, claiming that the most notorious spammers show no regard for any kind of anti-spam legislation, while legitimate marketers who would comply account for less than one-third of the spam people receive. Microsoft, AOL, and other direct marketers assert that a do-not-spam list would come with a hefty administration price tag, would be susceptible to hackers, and would probably limit the sales of DVD clubs and credit cards by legitimate firms. Anti-spam advocate and ePrivacy Group consultant Ray Everett-Church thinks implementing a do-not-spam registry is a good idea, even if more large companies than fly-by-night spammers comply, while Schumer contends that such a list would ease the prosecution of spammers in civil and criminal cases. Everett-Church thinks that a do-not-spam list would occupy "a comfortable middle ground" between the opt-in policies favored by the European Union and the opt-in approach preferred by U.S. lawmakers. On the other hand, Stanford Law School professor Lawrence Lessig thinks the registry would be ineffective, and proposes that commercial emailers should instead clearly label their messages as advertising. Meanwhile, certain consumer organizations would rather see the $75 million Schumer wants Congress to allocate for the anti-spam registry spent on other enforcement initiatives.
http://www.nytimes.com/2003/08/11/technology/11SPAM.html
(Access to this site is free; however, first-time visitors must register.)

Some U.S. states are having second thoughts about replacing their old voting machines with electronic systems because of a recent Johns Hopkins University report that calls the machines' security into question. Report co-author Avi Rubin, technical director of Johns Hopkins' Information Security Institute, studied a piece of the Diebold software source code accidentally left on a public Web site, and came to the conclusion that adolescents could fashion "smart cards" that would allow them to vote more than once; furthermore, the machines could be reprogrammed to alter election results by insiders. Rubin noted that reconstituting the actual vote is impossible because there is no paper ballot. The state of Maryland, which signed a $55.6 million contract for 11,000 touch-screen voting machines from Diebold, has asked an international computer security company to evaluate their security. Former electronic voting proponents such as The Leadership Conference on Civil Rights, which lobbied that new machines be deployed so that disabled people had an easy way to vote, now want President Bush and Congress to resolve the security issue. Other advocates, such as Riverside County, Calif., registrar of voters Mischelle Townsend, argue that electronic voting machines save significant sums in paper and have boosted voter turnout; Townsend adds that any attempts to tamper with such machines would be found out through intensive testing both before and after elections. Still, computer experts point out that there have been documented glitches with electronic voting systems: Some Diebold machines in Georgia registered votes for the wrong candidates, while a 7,000-vote error in Alabama was attributed to a system bug. Related issues computer scientists have raised alarms about include the clandestine testing the machines undergo before they are certified by the Federal Election Commission, and a 2001 General Accounting Office report stating that user friendliness and security are not high priorities in FEC testing standards.
Click Here to View Full Article

For more information about e-voting, visit http://www.acm.org/usacm.

Programmers and engineers are exerting their political influence online as a way to fight for civil liberties and other pressing issues. The leaders of these pushes have become cyber-activists because they have the money and the means to do so--Wes Boyd, for instance, amassed a fortune through the sale of screen savers and games, and used that capital to create MoveOn.org, a Web site originally designed as a platform for those opposed to the alleged political grandstanding surrounding former President Bill Clinton's impeachment. More recently, MoveOn was important in mobilizing many Americans opposed to the war against Iraq. DigitalConsumer.org was established in 2001 by Excite millionaires Joe Kraus and Graham Spencer as an outlet to battle legislation that seeks to grant entertainment companies extensive authorization to scout the Net for copyright infringement. The group's testimony was used by Congress to help stop a bill that would have waived entertainment companies' liability for any computers they damaged in their actions to curb digital song-swapping. Paulina Borsook, author of "Cyberselfish: A Critical Romp Through the Terribly Libertarian Culture of High-Tech," comments that the culture of Silicon Valley matured at a time when the Watergate incident and Vietnam engendered a distrust of government, but Doug Carlston of MoveOn points out that most programmers muffled their libertarian tendencies because activism was generally considered to be "unseemly" at the time. However, high-tech companies began to take a stronger political interest in the late 1990s, when legislation such as California's Proposition 211 would have made it easier to launch suits against companies with depressed stocks.
http://search.ft.com/search/article.html?id=030811001904

The development of rapid prototyping devices or 3D printers is about to advance with the development of machines that can print out moving components. Although RP devices can print out objects from metal, starch, plastic, or paper using a 3D blueprint as a guide, thus far the objects are cast as rigid blocks. A team of researchers at the University of California, Berkeley, hope to successfully print out a "flextronic" device with flexible joints and electronic elements within six months. UC Berkeley scientist Jeremy Risner wants to fabricate a fully functional device with both mechanical and electronic parts within 18 months, using a single printer equipped with multiple print heads. Various 3D printing methods such as selective laser sintering, laminated object manufacturing, and fused deposition modeling operate on the same basic principle: A printer head stamps out a fine powder while another ejects a bonding agent, forming a layer, and layers are deposited one on top of another. "RP is a mature industry, but the next move is to go into rapid manufacture, which requires material developments and increases in machine speed," notes Greg Gibbons of England's University of Warwick. Despite the challenges that lie ahead, RP has already significantly expedited industrial engineering, and sped up product development. Most notably, there is no configuration that is beyond an RP system's capabilities.
http://www.wired.com/news/technology/0,1282,59648,00.html

Researchers at the Institut d'Informatique et Mathematiques Appliquees de Grenoble have determined through the study of Wi-Fi network performance that a single person with a slow wireless link can reduce data transfer speeds for everyone using the same Wi-Fi access station. This phenomenon is directly related to the way the 802.11b standard's access protocol, Carrier Sense Multiple Access/Collision Avoidance (CSMA/CA), allots bandwidth to each user. In order to ensure that any user, regardless of individual access speed, can maintain a stable network connection, the CSMA/CA protocol throttles back everyone's maximum connection speed if just one person cannot link at that rate. This failsafe can result in the reduction of transfer speeds from 11 Mbps to approximately 1 Mbps. French researcher Andrzej Duda points out that users who do not use a large amount of bandwidth may not notice the slowdown, and says that some manufacturers have started to tackle the problem with their latest wireless networking hardware. Duda stresses, however, that this problem must not plague devices using 802.11g and other higher-bandwidth networking specifications. "In general, designers should be aware of this problem," comments Mingyan Liu of the University of Michigan. "But it would be worthwhile for researchers and engineers to come up with mitigating methods."
http://www.newscientist.com/news/news.jsp?id=ns99994018

The Centibots, whose development was funded by the Defense Advanced Research Projects Agency (DARPA), were put through their paces at the LinuxWorld Conference and Expo, where they were tasked to navigate a maze in order to locate a stuffed penguin. Several Centibots, working as a team and communicating through Linux, were able to accomplish this goal: A scout robot first mapped out the maze and its borders with a laser, and then two smaller robots were instructed to find the penguin via laptop; one of the two navigated the maze using the map, and scanned for the penguin, locating it within five minutes. SRI International developed the Centibots in conjunction with Stanford University, the University of Washington, and ActivMedia Robotics. The open-source Linux operating system was selected for the Centibots because it has become the software of choice for most scientists, and allows programs authored by other researchers to be borrowed and shared without the need for permission. IBM's Irving Wladawsky-Berger adds that "Linux has far more flexibility than anything else." SRI researcher Charlie Ortiz reports that DARPA is funding the Centibot project as a possible homeland security application called "urban surveillance." Robots programmed for urban surveillance would be able to navigate within and around buildings to locate hostages, for example. Distributed robotics initiatives such as the Centibot project seek to create robots that can work in teams, with DARPA's ultimate goal being a 100-robot swarm that can carry out rescue missions or detect chemical attacks.
Click Here to View Full Article

SCO has filed a $1 billion suit against IBM for allegedly including elements of its copyrighted Unix code in the free Linux operating system, and is threatening to sue all Linux users for as much as $700 each. Lee Gomes writes that SCO appears to have no legitimate intellectual property claims; moreover, SCO, in its Caldera incarnation, earnestly and freely distributed Linux software. Gomes compares SCO's current action "as if a magician gave away his secrets, then started suing his audience for learning how he did his tricks." The company refuses to show examples of this so-called infringement to outsiders unless they sign a Non-Disclosure Agreement, and Linux developer Ian Lance Taylor--one of the few who has seen such an example--reports that only about 80 lines of software out of 4.6 million came from Unix; furthermore, the code is not particularly important by itself, and can be easily rewritten. Gomes is puzzled that SCO will not reveal all the infringing code out of fear that Linux developers would replace it, which is logically what plaintiffs ought to demand in order to halt the legal injury they are suing for. Gomes concludes that this incident "is all just another inning in the common American business sport of 'patent shakedown,'" and anticipates that it will end with a dismissal similar to William R. Della Croce Jr.'s failed suit against Linux for allegedly infringing on the term "Linux." In fact, Gomes suggests that the SCO lawsuit could bolster Linux by pushing it to eliminate any arguable code and gaining a court-issued seal of copyright approval.

Neural-network technology, which is modeled after the synaptic architecture of the human brain, is being developed as a tool that enables computers to record information as patterns, and use those patterns to solve problems in much the same way human beings learn from experience. The technology does not follow the traditional programming paradigm, but instead relies on massively parallel networks that are taught to solve certain problems. Neural networking is generally considered to be the next step in the evolution of the computer industry, and the technology is being employed by corporations to anticipate customer behavior and detect fraud. Extracting meaning from data batches is migrating from the back office to the front office as such information becomes increasingly critical for decision-making, and data-mining technology derived from Predictive Modeling Markup Language and relational databases--two core neural-network elements--is being developed to fulfill this function. The most sophisticated neural-network and predictive-analysis systems currently implemented can predict future customer behavior using only a few dozen transactions by any given customer. The penetration of predictive analytics in the corporate marketplace was driven by the expansion of client-server computing and the increasing sophistication of networking protocols. Meanwhile, the Defense Advanced Research Projects Agency, the entity that initiated the original neural-network technology research, has allocated $29 million for a related project which seeks to automate routine military tasks by patterning the habits and preferences of battlefield commanders. Technology pioneer Robert Hecht-Nielsen calls the further development of neural-network technology "the most important scientific challenge of our time."
http://www.technewsworld.com/perl/story/31280.html

Legal expert Gerald Spindler warns that the open-source General Public License (GPL) is not valid in Germany, mostly because that country's law demands some type of liability or warranty from distributors or developers. Spindler wrote a 123-page study for the Verband der Softwareindustrie Deutschland, a industry group critical of open-source software with whom he claims no significant ties. Certain unalterable principles in German law, such as the need for liability and warranty, are not provided for under GPL, making that license not applicable in Germany. Spindler says the question of liability is complex and requires further work to determine whether employees, companies, or development participants share liability or whether one party is more liable than others; similarly, software that is based on previous open-source work is affected. Spindler explains that only in the past 18 months has the legal community focused on open-source software and its ramifications under law, and businesses might want to consider that no entity is clearly liable for the product. Although the German government has embraced open-source software, open source has not really had a chance to be properly addressed in the law. Doing so would entail writing the GPL in German and adapt it to both German and EU legal standards. The existing liability clause needs to be revisited to take into account the interests of users, developers, and competitors.
Click Here to View Full Article

In response to Vipul Prakash's observations and suggestions about spam control, Barry Shein finds fault with his position that there should be no per-message charging for email. Shein writes, "charges should be incurred to help pay for the resources being used and to inject some reality into decision-making about that usage," adding that a combination of statistical sampling and reasonable business relationships can help iron out hard usage problems that traditional network resource accounting is not equipped to handle. Shein notes that credible businesspeople realize how spam is hurting email, and believes that they could be persuaded to accept paying usage fees, provided the fee system is fair. Dave Crocker counters Shein's argument, claiming that usage-based charging would entail "astronomical" intrinsic costs, while setting up such a system would be a huge undertaking. Crocker reiterates Prakash's conclusion that the spam dialogue is overlooking three major issues: Every spam control solution is inherently limited and is not an end-all panacea for the spread of junk email; anti-spam mechanisms are likely to be more effective when combined; and email has accrued key features in the past three decades that people do not wish to give up. Crocker recommends caution, lest any changes to the email model, ostensibly to control spam, devalue Internet messaging.
http://www.technologyreview.com/articles/dialog0703.asp

The World Wide Web Consortium (W3C) has finally published its specification for creating Web forms. Initially scheduled to be released in March, XForms 1.0 uses Extensible Markup Language (XML) to bring greater flexibility to current documents based on Hyper Text Markup Language (HTML). According to a new draft from the W3C, "HTML forms have formed the backbone of the e-commerce revolution, and having shown their worth, have also indicated numerous ways they could be improved." Novell and IBM are among the companies that already are implementing XForms, which work with other tools recommended by the W3C and other devices and applications. The XForms standard calls for the forms to separate the code for defining Web page controls from the data they collect, which means XForms documents can be written once and used in many other places. W3C also says the ability to change XForms on the fly according to the information entered means there will be a decline in round-trips to the server. Disinterest from the market and vendors delayed the publishing of XForms as a recommendation.
http://news.com.com/2100-1013-5059904.html

Grace, the robot built in response to a general challenge by the American Society for Artificial Intelligence, is now preparing for more ambitious missions. Formally named the Graduate Robot Attending Conference in Edmonton, Grace successfully registered at the 18th annual National Conference on Artificial Intelligence in Canada, found her way to the meeting room, and then proceeded to deliver a prepared speech. She is a barrel-chested robot with no arms or legs, and moves on wheels guided by sensors near the ground. The robot features a flat panel computer screen that displays Grace's animated face, as well as a camera, speech synthesizer, and a microphone. The five research teams that contributed to Grace aim to improve her performance at the upcoming International Joint Conference on Artificial Intelligence in Acapulco, Mexico, from Aug. 9-15. Reid Simmons of the Carnegie Mellon University Robotics Institute, leads the team that designed Grace's software and hardware structure. He says Grace previously did in one hour what a person would take 10 minutes to do. The goal this time is to improve her performance by allowing her to do multiple tasks at once--finding the end of the registration line while en route, for example. In Canada, Grace embarrassed her creators when she cut in line at the registration table. In addition to improved speed and movement, the Grace team also intends to bolster her conversational skills so that she can engage in more natural exchanges with human attendees. Grace will be accompanied in Mexico by George, her male counterpart who is identical except for voice and display; in the future, the researchers plan to make the two robots collaborate to finish tasks faster.
Click Here to View Full Article

The vulnerability of the Internet is a lingering problem, and security experts concur that private industry must lead the charge to improve Net security. "Internet security is...about teaching people that there is a lot more to security than buying a cable modem router with a firewall built into it," posits Brian King of the CERT Coordination Center at Carnegie Mellon University's Software Engineering Institute. King says the problem is not innovative network intrusion methods so much as a lack of security training among computer scientists as well as vendor laziness; for Internet security to improve, vendors and administrators need to devote more attention to the problem and take rectifying steps--administrators must keep up with best practices for shielding against specific intrusions, while vendors must concentrate on delivering secure rather than feature-rich software. Other strategies King advises vendors to follow include regularly upgrading operating systems with the latest patches and keeping antivirus software up to date. Peter Allor of Internet Security Systems says that installing patches and keeping abreast of reported security flaws is not a top priority among enterprises, while security-solutions vendors need to design products with security in mind. Internet Security Alliance (ISA) COO Larry Clinton says that the National Strategy to Secure Cyberspace clearly states that the public and private sectors must work together to improve Internet security, and notes that the federal government plans to set up a private communications network for exchanging information on security threats. Clinton adds that one way the government could make vendors more security-conscious would be to require suppliers to comply with security standards. Meanwhile, the ISA is attempting to create an incentive system to encourage vendors to bolster Internet security.
http://www.newsfactor.com/perl/story/22035.html

Some hearing-impaired people have conflicting feelings about technology designed to translate American Sign Language into spoken and written speech, the latest example being Jose Hernandez-Rebollar's AcceleGlove, a sensor-laden glove that converts hand and arm movements into vocalizations or text messages. Previous glove technologies take a long time to spell out words and have a small vocabulary. The AcceleGlove can translate almost 200 words and a few simple phrases, and can understand both the alphabet and dynamic gestures. George Washington University doctoral student Hernandez-Rebollar notes that a two-glove system will be needed to allow the wearer to communicate the entire ASL vocabulary, while a preinstalled dictionary would expand the range of gestures the glove can translate. However, fellow glove translator inventor Ryan Patterson of the University of Colorado observes that the technology is limited because it cannot take facial expressions into account. American Sign Language Institute director Paul Mitchell says that deaf people may be resentful of such technology, not just because of its limited vocabulary, but because it goes against their own cultural view that deafness is a unique trait rather than a disability. Certain organizations believe that imposing such technology on deaf people as a "cure" for their condition would force the hearing-impaired to radically alter their lives at the behest of the hearing world, rather than let the hearing community accommodate them.
http://www.wired.com/news/technology/0,1282,59912,00.html

German researchers have developed a 3D graphics program that can help forensics experts in their efforts to reconstruct faces from the skeletal remains of dead people found by the police. The graphics system of Kolja Kahler and Jorg Haber of the Max Planck Institute for Computer Science in Saarbrucken, Germany, and their colleagues speeds up the traditional forensic process of having an artist make a plaster cast of a skull and cover it with clay to mimic flesh. This laborious process tends to take the forensic artist several weeks. In addition to speeding up the facial reconstruction process, the program's animation approach is beneficial in that it allows police to add expressions to the facial reconstruction with the click of a mouse. "An artist cannot give the model personality, and a lot of times, that's what people recognize," says Todd Matthews of Doe Network, a global organization that works with forensic artists to match unidentified bodies. Police typically have forensic artists create one sculpture, and the lack of facial expressions often makes the model difficult to recognize. Dieter Buhman, a forensics expert at the Institute of Forensic Medicine at Saarland University in Saarbrucken, has recommended that the team create a way for the police to add hair color, hairstyles, and skin texture to reconstructed faces using their 3D graphics program.
http://www.newscientist.com/news/news.jsp?id=ns99994005

Lauren Weinstein was on the receiving end of the first spam ever--an email from Digital Equipment--in May,1978, while involved in work to develop an early version of the Internet. Today, junk email accounts for 49 percent of network traffic, according to Brightmail, a manufacturer of anti-spam software. Although efforts to fight spam are starting to gain momentum, such as lawsuits, legislation, and technology, an approach that combines these strategies is what is needed to seriously have an impact on the proliferation of unsolicited email, write Lorraine Woellert and Stephen H. Wildstrom. The first step to take in eliminating spam is to strengthen laws to allow for the establishment of a Do-Not-Spam registry that would give people the choice of whether they want to receive commercial email. After beefing up the books, authorities must do a better job of expanding the policing effort, such as by allowing users to sue spammers directly with minimum statutory damages of about $100 per message. U.S. officials will need to get other countries involved in strengthening their laws and enforcement to keep spammers from moving offshore. The tech industry will need to develop new rules for transmitting messages so that spammers can no longer fake their identities. Also, consumers and companies will need to invest in good filtering tools.
Click Here to View Full Article

SCO Group's $3 billion lawsuit against IBM for allegedly incorporating patented Unix code into Linux has ignited a flurry of speculation on whether the open-source development process is in need of serious revision. Some IT analysts are advising companies to be circumspect when they consider adopting Linux, especially with SCO threatening litigation against users running the Linux system; meanwhile, some Linux developers admit that they must be more thorough in ensuring that the code they write is original. This view is not shared by Open Source Development Lab CEO Stuart Cohen, who insists that his lab adheres to "a very structured" Linux development process. He notes, for example, that all code contributions must be approved by Linux founder Linus Torvalds, Linux kernel supporter Andrew Morton, and end users. Though Torvalds acknowledges that the open-source development process cannot guarantee against the inclusion of patented code, he argues that the responsibility of sifting through patents for potential legal problems should be left to lawyers, not engineers. Stuart Meyer of Fenwick & West thinks open-source developers should be more "cognizant" of patent and intellectual property issues, and evaluate their work to prevent infringement. Some observers say that indemnification should be extended beyond specific software environments and operating systems, while Microsoft senior VP Eric Rudder suggests that new tools such as digital-rights management can be deployed to prevent unauthorized code copying. Changes to the open-source development process will not occur until enough participants agree that such a strategy is a wise and necessary move.
Click Here to View Full Article