Timely Topics for IT Professionals
About ACM TechNews
ACM TechNews is published every week on Monday, Wednesday, and Friday.
ACM TechNews is intended as an objective news digest for busy IT Professionals. Views expressed are not necessarily those of either HP or ACM.
To send comments, please write to email@example.com.
Volume 5, Issue 526: Wednesday, July 30, 2003
- "Big Bang Meets Graphics at ACM Confab"
CNet (07/28/03); Kanellos, Michael
Among the scientific research that computer graphics technology is contributing to is a Cambridge University project that compares ancient radiation patterns to computer simulations to determine that the universe has a planar shape, according to Cambridge professor Anthony Lasenby, a keynote speaker at the ACM SIGGRAPH conference this week in San Diego. Stanford University professor and recipient of ACM's 2003 Steven Anson Coons Award for Outstanding Creative Contributions to Computer Graphics Pat Hanrahan observed that the simulation of light patterns and physical movement in natural phenomena has become an attainable goal for scientists thanks to the increasing calculative power of computers. Jim Blinn of Microsoft Research noted that some researchers are investigating ways to increase the realism of computer simulations by restructuring computer graphics, which are currently generated by segmenting objects and surfaces into triangles. Blinn is focused on devising a method to enable modern hardware to more easily carry out smooth-surface calculations, and this technique may eventually be integrated with traditional polygonal rendering. Many exhibitors and speakers at this year's SIGGRAPH hail from Hollywood special effects companies such as Pixar Animation and Industrial Light and Magic. Research institutions represented at the conference, in addition to Cambridge, include Harvard, Purdue, the Rochester Institute of Technology, and the University of Southern California.
- "Federal Data Searches on Hit List"
Wired News (07/29/03); Singel, Ryan
Sen. Ron Wyden (D-Ore.) is working with a wide range of privacy groups to restrict data mining of federal and commercial databases by the CIA and the Defense, Homeland Security, and Treasury departments. His newly drafted Citizens' Protection in Federal Databases Act would prohibit agents from fishing in databases for people who fit a particular profile, instead allowing them to only search for information on already-identified suspects. In addition, any federal agency that did not report to Congress about new database programs within 60 days would have its funding suspended. Wyden already succeeded in freezing the Pentagon's Terrorism Information Awareness (TIA) program by attaching provisions to a Defense Department spending bill. According to records won by lawsuit, the Electronic Privacy Information Center (EPIC) says a limited number of FBI offices searched public records--such as credit bureaus, LexisNexis, Westlaw, and ChoicePoint--67,932 times in 2000. Extrapolating from that information, all FBI offices made a total of 1,698,300 public records searches in that same year, searches the agency says contributed to a number of captured fugitives. Still, EPIC's David Sobel says the public is not ready to sacrifice privacy yet for increased security, even after the Sept. 11 terrorist attacks. He says Wyden's efforts have strong public backing, as evidenced by the backlash against TIA and CAPPS II, a passenger flight screening system.
- "Race Is On for Mobile Graphics"
EBN (07/28/03); Merritt, Rick
One of the major events at the ACM SIGGRAPH conference will be the official debut of OpenGL Embedded Systems version 1.0, marking the formal start of a race to define handheld graphics. Nokia and Motorola will state their support for OpenGL ES at the conference; Nokia chose the standard for 3D in a Java environment while Motorola is incorporating OpenGL ES in its mobile offerings and chips for cellular phones, telematics, and set-top boxes. Meanwhile, Sun Microsystems is expected to formally announce itself a member of the Khronos Group, the industry alliance defining OpenGL ES. Khronos Chairman Neil Trevett declared that 3D gaming for handsets will be a primary driver of mobile graphics, and he expects OpenGL ES hardware accelerators to ship by year's end. Trevett added that the majority of 2003 OpenGL ES shipments will be software deployments that utilize fixed-point graphics, while the forthcoming accelerators will deliver floating-point graphics. Microsoft is readying the inaugural draft of its Direct3D Mobile handheld graphics application-programming interface (API), while Palm is preparing version 6.0 of its operating system for personal digital assistants and cell phones, which will boast a graphics architecture that integrates design concepts from OpenGL and Postscript. George Hoffman of Palm's OS group said the operating system would not include Direct3D Mobile support, while OpenGL ES support will also be missing because of certain design disparities. Support for Palm's mobile graphics architecture is being implemented with collaborating chipmakers ATI Technologies, Motorola, Texas Instruments, Intel, and MediaQ. SIGGRAPH 2003 chair Alyn Rockwood proclaimed that the transition to mobile graphics will be a key theme of the conference, while Ericsson researchers will present a paper citing mobile phones' need for greater graphics resolution than PCs.
Click Here to View Full Article
- "SGI Unveils Update to OpenGL Graphics Spec"
TechNewsWorld (07/28/03); Lyman, Jay
SGI declared that the OpenGL 1.5 application programming interface (API) introduced on July 28 will enable developers to write shader applications in real time and facilitate more realistic graphical image rendering. The update features OpenGL shading language that can be used by any graphics card regardless of vendor or platform, while Shawn Underwood of SGI's Visual Systems Group expects the enhanced standard to "fundamentally change the industry." Mercury Research President Dean McCarron explains that OpenGL 1.5 will boost the realism of workstation graphics, and enable the OpenGL API to consistently compete with Microsoft's DirectX graphics interface. He also notes that new hardware is forthcoming that will supply enough graphics power to support new 3D rendering proficiency, allow programs to run on the graphics chip instead of the processor, and generate more authentic-looking surfaces and textures. OpenGL 1.5's shading capabilities are being showcased at the ACM SIGGRAPH conference in San Diego, but McCarron estimates that programs that take advantage of such applications will not penetrate the mass market until 2004. "With the addition of an innovative shading language, software developers now have high-level access to the programmable features of modern [graphics processing units]," proclaims Kurt Akeley of Nvidia. SGI also announced a partnership with Sun Microsystems to build Java support into the OpenGL API.
- "And Now, Here Comes 'Spray-On Electronics'"
International Herald Tribune (07/25/03); Schenker, Jennifer L.
IBM, Philips, and Bell Labs are developing organic transistors as a basis for plastic electronics, but Plastic Logic of Cambridge, England, claims to have a competitive advantage with a patented technique for printing "spray-on" polymer-based circuits. Although Plastic Logic chips do not boast the speed of crystalline silicon chips, Plastic Logic co-founder Richard Friend says they are sufficiently cheap and fast enough to supplant silicon in updateable active matrix displays incorporated into electronic signs and electronic newspapers. Plastic electronics promise to eliminate the need for manufacturing chips in sterile "clean rooms" and the associated costs, and make ubiquitous electronics an achievable goal. Plastic electronics could be the keystone of a range of new products and applications, including flexible e-paper and e-textiles, disposable electronics, advanced biosensors, electronic labels, intelligent packaging, and roll-up displays for mobile phones. One of Plastic Logic's partners, Cambridge Display Technologies, is developing organic light-emitting displays as a complementary technology for plastic electronics. "The flexibility that plastic electronics offers allows us to enter new markets and enter new product lines," declares Jim Welch of Gyricon. Plastic electronics' initial applications are expected to be in retail stores, probably as price displays that can be remotely updated via computer. More advanced products such as changeable e-books will not emerge until wireless technologies proliferate and the cost of display materials significantly declines.
- "Robert V.D. Campbell, ACM Co-Founder, Dies at 86"
The Sudbury Town Crier (7/9/03)
Robert van Duyne Campbell, one of the eight founders of the Association for Computing Machinery, died July 1 in Boston at the age of 86. The noted computer pioneer collaborated with Howard Aiken and IBM to develop the first programs for the Mark I calculator. After his discharge from the Navy in 1946, he returned to Harvard and contributed to the design of the Mark II computer. It was his suggestion that the Mark II use the now standard floating point approach for number representation. Campbell later joined the Research Center at Burroughs Corporation and in 1966 joined MITRE Corporation as a systems analyst. His ACM contributions began with a simple notice distributed in late June 1947 among colleagues known for their interest in computers. "After making some inquiries during May and June, we believe there is ample interest to start an information association of many of those interested in the new machinery for computing and reasoning. Since there has to be a beginning, we are acting as a temporary committee to start such an association," the notice stated, and it was signed E.C. Berkeley, R.V.D. Campbell, J.H. Curtiss, H.E. Goheen, J.W. Mauchly, T.K. Sharpless, R. Taylor, and C.B. Tompkins. Campbell, in fact, served as ACM's first Treasurer. He is survived by his wife of 64 years, Winifred, a son, two daughters, and three grandchildren.
Click Here to View Full Article
- "World's First Tunable 'Photon Copier' On a Chip"
Researchers at the University of California at Santa Barbara (UCSB) have moved one step closer to an all-optical network by constructing an integrated photonic circuit that can transcribe data from one optical hue to another by embedding a widely tunable laser and an all-optical wavelength converter into a single indium phosphide chip. Switching between colors is key to successfully transmitting data along a multi-node fiber-optic network, and the process currently involves converting photons to electrons. The UCSB breakthrough, which was financed by the Defense Advanced Research Projects Agency's Microsystems Technology Office, eliminates the need for an electronic bridge between the two colors because the laser and converter are no longer on separate chips. The device also offers considerable performance gains and maintains signal quality. The tunable laser, known as a Sampled-Grating Distributed-Bragg-Reflector laser, was invented by Larry Coldren, who led the UCSB research team along with electrical and computer engineering professor Daniel Blumenthal. The converter component is an SOA-based Mach-Zehnder interferometer that incorporates signal regeneration, which enables the device to sustain or improve image quality. A paper presented by UCSB grad student and research team member Milan Masanovic at the 15th annual meeting of the Indium Phosphide and Related Materials Conference detailed the integrated photonic circuit, describing the device as "a critical step towards realizing truly optical switches and networks."
Click Here to View Full Article
- "AI Depends on Your Point of View"
Wired News (07/29/03); Shachtman, Noah
The Information Processing Technology Office (IPTO) of the Defense Advanced Research Projects Agency (DARPA) has launched an effort to develop computers that can think for themselves, and the Real-World Reasoning project is part of this effort. The project seeks to give computers the ability to study situations from multiple angles and learn from experience, possibly through the integration of straight-up logic, probabilistic reasoning, game theory, and strategic thinking. Human beings do not merely feed new information into a database, notes IPTO chief Ron Brachman. He says that "[the data's] got to jive with what we know already. Or we've got [to] adjust our previous understanding." One technique people use to facilitate that adjustment is to look at the situation from a different context. It is doubtful that the Real-World Reasoning project will yield computers that exhibit the same mental flexibility as people, although it is hoped that the program will improve their mode of reasoning.
Click Here to View Full Article
- "U.S. Shrugs Off World's Address Shortage"
CNet (07/28/03); Charny, Ben
Outside the United States, the world is running out of IP addresses under the IPv4 system and is adopting Internet Protocol version 6 (IPv6). North America was assigned 70 percent of the 4.3 billion possible IP addresses under the current v4 system, so does not feel the pressing need to create more room. In Asia, however, analysts say that quick adoption by Chinese and Indian users and network-connected Koreans will soon deplete their available pool of original addresses. The Defense Department is pioneering the U.S. transition to IPv6 because of that standard's increased security and network management capabilities, and expects to migrate all its networks by 2008. IPv6 would allow virtual private networks to more easily support vast numbers of connected devices, and would also solve always-on problems faced by network administrators; but getting all network devices to rely on IPv6 would require a much greater push than the preparation for Y2K, and NTT product engineering director Cody Christman says many firms simply lack the motivation, especially since the two standards are compatible. NTT's DoCoMo wireless carrier has already switched to IPv6 out of fears that its Web-enabled mobile phone subscribers would run out of addresses. IPv6 quadruples the length of the address from 32 bits to 128 bits, and thus makes room for 1,000 IP addresses per person on the planet. Experts say that U.S. companies should not become complacent if they intend to continue their lead in innovation, especially since emerging technologies such as voice over IP and radio-frequency ID tags will increase demand for individual addresses.
- "Symposium Connects Students With Latest Tech Advances"
SiliconValley.com (07/28/03); Sweeney, Frank
Some 1,500 high school students 15 to 18 years old have convened in San Jose for the National Youth Leadership Forum on Technology, a 10-day symposium featuring workshops, seminar, campus visits, roundtable discussions, and demonstrations of Silicon Valley's latest technological advances. The attendees, who hail from across the United States, score high in academic achievement, in the A to B+ range, according to Brett West, a forum staff member. One attending student, Tyler Wolf of Great Falls, Mont., says that he came to the forum "to see what's new and interesting and to get my feet in the water." Among the technologies being demonstrated to students is an Intel Itanium computer that uses dual projectors to transmit an image onto a screen that viewers can see in three dimensions using traditional 3D glasses. West comments that forum attendees have an interest in technology, science, and graphics design, while some are even fledgling entrepreneurs. He adds that students will be asked to choose a problem and develop a high-tech solution. One week later, their ideas will be demonstrated at a trade show-like session.
Click Here to View Full Article
- "Experiment to Outfit Classroom With Sensors"
Daily Bruin (07/28/03); Mignot, Christian
UCLA electrical engineering professor Mani Srivastava is leading a project to equip a first grade classroom with minuscule electronic sensors in order to monitor student interactions and how they relate to academic performance. The National Science Foundation has committed $1.8 million toward the experiment because of the insights the project could yield about educational methods, software and hardware applications, and children's speech development. Srivastava explains that students will wear sensor-studded caps that will keep tabs on location and head orientation, as well as record their speech with microphones. Researchers will study how students carry out the teacher's instructions through the use of sensor-outfitted puzzle pieces and other objects, and task tables equipped with magnetic systems. The classroom itself will also boast microphones and cameras to watch students; the microphones will capture sound clips that researchers can analyze to study children's speech patterns, according to Srivastava. Gregory Chung of the UCLA Graduate School of Education and Information Studies' Center for the Study of Evaluation notes that the sensors would help teachers appraise individual student performance in order to see what kinds of problems students may face. The data collated by these various systems goes through a central computer running a software program created by computer science professor Richard Muntz that sports data-mining capabilities. Srivastava says that "The use of sensors in this manner will allow people to talk and interact with the physical world."
Click Here to View Full Article
- "Open Source Gets Down to Business"
Technology Review (07/25/03); Stenger, Brad
Former Wall Street IT director and MIT engineering graduate Robert Lefkowitz now advises the open-source community on how its products can better meet the needs of the commercial sector. Most recently, he spoke at a Linux summit in Ottawa, Canada, and at the Open Source Convention in Portland. He says the most attractive benefit of open-source technology in business is not the free licensing, but the ability to vet code upfront as opposed to hiring a programmer who looks promising. Open-source technology lures people from the commercial world with its promise of continued learning and is sort of a middle ground where students and current practitioners can intermingle and share ideas. Lefkowitz says open-source software has two development drives, one being academic and more expositional in nature, and the other being commercial and therefore focused on scalability and performance. Smaller IT groups do not have the resources to test code on many different platforms but testing and porting work is made much simpler with open-source software, Lefkowitz adds. Larger IT organizations also have incentive to adopt open-source software because they can use it as a base for an innovative proprietary product, such as Microsoft did with ActiveDirectory. In general, open source is good for business because it fosters the type of collaboration and idea-sharing valued in corporations today.
Click Here to View Full Article
- "CERT Project Takes Cue From National Security Plan"
Computer Business Review (07/28/03); Murphy, Kevin
The CERT Coordination Center is looking to create information sharing best practices that could be used in a security Information Sharing and Analysis Center (ISAC), a recommendation included in the U.S. National Strategy to Secure Cyberspace that was published this year. CERT has created the Cyber Security Information Sharing Project (CSISP) to head the effort, which ultimately could help coordinate private networks in the United States to protect against Internet attacks. ArcSight, which makes security event management software, will contribute free products to the effort. Three universities will deploy ArcSight software, which will capture security events from logs of firewalls and other devices, and analyze the data. In addition to investigating information sharing best practices, the two-year CSISP project will seek to refine the Intrusion Detection Message Exchange Format (IDMEF) and Incident Object Description and Exchange Format (IODEF) standards, out of the Internet Engineering Task Force.
Click Here to View Full Article
- "Helping Machines Think Different"
Wired News (07/29/03); Shachtman, Noah
The Defense Advanced Research Projects Agency (DARPA) says it has embarked on a series of projects in recent months that on the surface may seem isolated, but are in fact part of an overarching push to make computers capable of thinking for themselves. LifeLog, the most well-known of these projects, seeks to track all aspects of a person's life and feed this information into a database; a program is supposed to extract narrative threads from this data to deduce that person's relationships, experiences, and traits. "Our ultimate goal is to build a new generation of computer systems that are substantially more robust, secure, helpful, long-lasting and adaptive to their users and tasks," explains American Association for Artificial Intelligence President Ron Brachman, who was recently appointed head of DARPA's Information Processing Technology Office (IPTO). "These systems will need to reason, learn and respond intelligently to things they've never encountered before." Brachman insists that LifeLog is not a profiling or terrorist-tracking tool, but an episodic-memory system forming the basis of a sophisticated electronic assistant that intuits the habits and preferences of its boss. IPTO's $29 million Perceptive System that Learns (PAL) program is an effort to build self-improving software that references episodic memory to automate the scheduling of meetings and other tasks. Brachman observed in a recent presentation that the growing complexity of computer systems increases their fragility and vulnerability to attack. So that a computer can learn and adapt to such factors, it must build a catalog of existence the same way people do.
- "The Future of Human Knowledge: The Semantic Web"
TechNewsWorld (07/28/03); Koprowski, Gene J.
An international team of scientists is defining and devising standards, protocols, and technologies that will form the foundation of the Semantic Web, a more context-aware version of the Internet envisioned as an online source for the collective scientific, business, and artistic knowledge of the human race. The Semantic Web, which optimistic forecasters believe is just a few years away, will be able to conduct searches based on ordinary language rather than keywords. The project is being spearheaded by the World Wide Web Consortium; participants include researchers at computer companies and major academic institutions such as Kyoto University, Stanford University, and MIT, while the Pentagon's Defense Advanced Research Projects Agency (DARPA) is a Semantic Web underwriter. Maturing technologies thought to be the strongest candidates for implementing the Semantic Web include DARPA agent markup language, resource description framework (RDF), and Web ontology language (OWL). Many businesses are too impatient to wait for the Semantic Web to emerge and are investigating other stop-gap solutions, such as the automation of routine e-commerce tasks by software agents. Meanwhile, niche search engines that seek out data based on user preferences are being fleshed out by Penn State researchers and others. Extreme Logic strategic consultant Steve Woods notes that analysts expect the Semantic Web to supplant the current Web within two years, although his firm considers 2010 to be a more realistic projection for such a development.
- "Australia Considers Sending Spammers to Jail"
ZDNet Australia (07/25/03); Pearce, James
Symantec managing director John Donovan is a member of an advisory group helping develop new Australian anti-spam legislation, but thinks that the penalties should be more severe than the final document states, noting that companies in the United States fined for spamming consider the fines part of their operating expenses. He believes that the government has gone for speedy introduction of the legislation rather than stronger penalties, and that criminal penalties may be introduced later. Donovan notes that most spam comes in from other nations and that the legislation will not stop spam by itself. He hopes also that other nations will adopt it. "What [the legislation is] really aiming at is list-harvesting," Donavan comments.
- "DNS Is Busting Out All Over"
Network World (07/28/03) Vol. 20, No. 30, P. 1; Fontana, John
The Domain Name System is the basis of several new technologies, showing that the 20-year-old system is still robust and unlikely to fade away quietly. Radio frequency identification (RFID) technology is being tied to DNS as a way to keep data off the chip and on the network. The Auto-ID Center group at MIT first came up with the idea as a way to reduce the size and cost of each RFID tag, since product data is stored on the Internet rather than on silicon; through the Auto-ID Center's Object Naming Service, the product code stored on the RFID is translated into a domain link on DNS and then automatically downloads the product data. Enum, or Electronic Numbering, is another new DNS application that moves phone accounts from the traditional telecommunications system to Internet addresses. People can be reached at just one Internet address, which will route the message to phone, email, or other Internet applications. DNS creator Paul Mockapetris says the new uses for DNS show that its utility has not yet peaked, but rather that it is useful for managing higher-level services, not just hosts. However, the explosion of DNS data will require management tools that simplify DNS maintenance: "Originally we built DNS so it was very simple to configure, but everybody was a computer scientist," Mockapetris says. In addition, almost 70 percent of corporate DNS servers are improperly configured, an issue of increasing importance as more real-time services rely on the system. Mockapetris predicts DNS will also contribute to greater Internet security by facilitating widespread use of public key infrastructure. Mockapetris says, "The new challenge is to see whether we can continue to use DNS for whatever we like...You want clean names so to speak, and clean data, it's what applications need to live on."
- "Stuff That Works"
Canadian Business (07/21/03) Vol. 76, No. 13, P. 28; Holloway, Andy; Wahl, Andrew
North American companies are embracing three technologies--Wi-Fi, voice over Internet Protocol (VoIP), and the open-source Linux operating system--that may actually fulfill their potential and transform business models. Wi-Fi promises anytime, anywhere Internet access via high-speed wireless broadband; American and Canadian companies and consortia such as Cometa Networks and Verizon are scrambling to set up Wi-Fi hotspots, and independent and regional ISPs have also hopped on the Wi-Fi bandwagon. In addition, computer manufacturers expect to reinvigorate sales with Wi-Fi, while security issues about the technology should be ironed out with the forthcoming 802.11g and 802.11i releases. Lingering concerns about how much people should be billed for Wi-Fi access, its harmonization with next-generation mobile phone networks, etc., should be allayed over the next few years. VoIP--the convergence of voice and data networks--is proving popular because it dramatically cuts long-term support and long-distance costs, and International Data (IDC) predicts that the worldwide market for IP telephony enterprise equipment will more than triple between 2003 and 2007. VoIP's advantages in the area of call-center routing and customer inquiry management helped make government, education, health care, service-oriented financial, and retail companies primary adopters, while extension mobility is bringing in even more VoIP users. Companies such as Google and even governments are considering or relying heavily on Linux, which allows people to access a treasure trove of cheap software. Linux shipments are expected to increase 100 percent by 2007, and George H. Young International's Nigel Fortlage declares that Linux "really is becoming very mainstream and, certainly to us, quite valuable."
Click Here to View Full article
- "A More User-Friendly Direction"
Darwin (07/03); Shneiderman, Ben
Computer users are often frustrated by unreliable software and interfaces, and improving product usability requires detailed documentation of the extent of the problem, documentation that is in short supply. In his book, "Leonardo's Laptop: Human Needs and the New Computing Technologies," Ben Shneiderman notes that monitoring users and logging usage would yield accurate data on user application preferences, and provide feedback that product makers could use to improve designs. He argues that consumers need to apply more pressure on companies to develop better-quality products and make quality an overriding priority. Shneiderman explains that technology providers need to treat consumers and their opinions with respect, while consumers could lobby for government entities and independent consumer organizations to carry out detailed assessments of products and user experiences. "Consumers should be aware of the need to promote their rights against proposals that would allow software suppliers to avoid liability for damages, change license terms at any time, prohibit negative reviews and access your files so that they could remove software if they determine you have violated the license," Shneiderman writes, encouraging users to air their grievances to suppliers, government officials, and industry groups. The author puts forward the idea that users should be financially compensated by suppliers for the problems they experience, and goes on to say that the reward system cuts both ways: Programmers and software engineers should also receive recognition for more reliable products. Software providers often claim that such solutions are too costly, but Shneiderman counters that the cumulative loss in production due to software glitches far outweighs the cost of implementing measures to boost software quality. A major challenge for deploying the new computing paradigm is to overcome the cynicism typical of the computing industry and the tendency to prioritize technology over user needs.
- "Hybrid Semiconductor-Molecular Nanoelectronics"
Industrial Physicist (07/03) Vol. 9, No. 3, P. 20; Likharev, Konstantin
The creation of denser integrated circuits based on complementary metal oxide semiconductor (CMOS) technology is limited by the sensitivity of silicon field-effect transistor parameters to random alterations of device size, and many engineers and physicists agree that the solution to the problem lies in hybrid semiconductor-molecular circuits that merge the advantages of nanoscale elements with the benefits of patterning methods. Single-electron devices offer the strongest prospects for single-molecule components because the device-to-electrode connections do not need to be highly conductive in order for the operation mechanism to function. Challenges researchers need to meet include extending chemically directed self-assembly beyond single-layer growth on smooth wafers to the reliable deposition of three-terminal molecules on nanowire structures; a team from the State University of New York (SUNY) at Stony Brook is striving to accomplish this via a molecule that may operate as a latching switch directed by two input signals. Such a device, with CMOS layer transistors, has been arranged into distributed crossbar neuromorphic networks, or CrossNets, which consist of a grid of tiles occasionally interrupted by contacts to semiconductor cells. CrossNets promise a high degree of connectivity for such operations as image recognition, and the SUNY researchers estimate that a CrossNet circuit with a large but sufficient power of 10 Watts per square centimeter may furnish an intercell communication latency of roughly 200 nanoseconds, making the device five times faster than the human brain. Neuromorphic networks can only carry out their tasks through training, and back-propagation and other well-developed neural training techniques are inapplicable to CrossNets. The SUNY group has enjoyed a measure of success in image recognition and plans to refine the CrossNet for use in image classification, a breakthrough that could revolutionize pattern analysis; this work could conceivably lead to a "self-evolving" computer system capable of self-awareness and reasoning.