Association for Computing Machinery
Welcome to the May 27, 2011 edition of ACM TechNews, providing timely information for IT professionals three times a week.

Please note: In observance of the Memorial Day holiday, TechNews will not be published on Monday, May 30. Publication will resume Wednesday, June 1.

Also, please download our new ACM TechNews iPhone App from the iTunes Store by clicking here and our new ACM TechNews iPad App by clicking here.

HEADLINES AT A GLANCE


Code-Cracking Machine Returned to Life
BBC News (05/27/11) Mark Ward

England's National Museum of Computing recently finished restoring a Tunny machine, which was a key part of Allied code-cracking during World War II. The restoration effort began in 2005 and was led by computer conservationists John Pether and John Whetter. "As far as I know there were no original circuit diagrams left," Pether says. "All we had was a few circuit elements drawn up from memory by engineers who worked on the original." He says the most difficult part of the rebuilding project was getting the six timing circuits of the machine working in unison. The restoration project was able to succeed because the machines were built by the Post Office's research lab at Dollis Hill, and all of the parts were normally used to build telephone exchanges after the war, according to Whetter. The first Tunny machine was built by Bill Tutte in 1942 and worked alongside the early Colossus computer to unscramble messages sent by the Nazis. After the war the 12-15 Tunny machines were dismantled and recycled for scrap. "We have a great deal of admiration for Bill Tutte and those original engineers," Whetter says. "There were no standard drawings they could put together. It was all original thought and it was incredible what they achieved."


Cybersecurity Plan Faulted
Wall Street Journal (05/27/11) Siobhan Gorman

The U.S. Chamber of Commerce rejected the Obama administration's core proposal for enhancing U.S. cybersecurity systems, claiming the plan has significant areas of concern. The Chamber's main concern is that the White House plan would require certain companies running the most important components of cyberinfrastructure to submit to more intense outside oversight of their cybersecurity practices. The president's proposal involves the Department of Homeland Security and industry representatives identifying the companies that have the most influence on U.S. cyberinfrastructure, developing plans for those companies to address cyberthreats, and having an outside auditor review the plans. "Layering new regulations on critical infrastructure will harm public-private partnerships, cost industry substantial sums, and not necessarily improve national security," the Chamber says. White House officials say that a national cybersecurity plan is crucial to protecting the country's critical infrastructure, and they believed they had crafted a flexible, business-friendly plan. "Our proposal strikes a critical balance between strengthening security, preserving privacy and civil-liberties protections, and fostering continued economic growth," says White House spokesman Nicholas Shapiro.


IT Grads in Demand
NextGov.com (05/23/11) Brittany Ballenstedt

Employers are looking to hire graduates with degrees in computer science more than any other college graduates, according to a new survey from the National Association of Colleges and Employers (NACE). More than 56 percent of computer science majors who have applied for a job have received an offer, giving the major the highest offer rate for the class of 2011. NACE surveyed more than 50,000 college students and found that the job offer rate for computer science graduates has increased 13.8 percent from 2010. Half of engineering graduates who have applied for jobs have received an offer, and the overall job offer rate for 2011 graduates has risen to 41.2 percent from 38.2 percent a year ago. In a separate survey released earlier in the month, NACE reported that computer science graduates were receiving starting salaries of $63,000.


New Bandwidth Management Techniques Boost Operating Efficiency in Multi-Core Chips
NCSU News (05/25/11) Matt Shipman

North Carolina State University (NCSU) researchers have developed two techniques that make semiconductor bandwidth allocation and data prefetching more efficient. Combined, the techniques can increase the performance of multicore chips by as much as 40 percent, according to the researchers. "The first technique relies on criteria we developed to determine how much bandwidth should be allotted to each core on a chip," says NCSU professor Yan Solihin. The researchers use data from the hardware counters on each chip to determine which cores need more bandwidth. "The second technique relies on a set of criteria we developed for determining when prefetching will boost performance and should be utilized, as well as when prefetching would slow things down and should be avoided," Solihin says. The prefetching criteria enables chip developers to make multicore chips that work more efficiently by allowing individual cores to turn the prefetching ability on or off as needed.


Recapping the US Ignite Gigabit Applications Workshop
Computing Community Consortium (05/24/11) Erwin Gianchandani

The recent US Ignite Gigabit Applications Workshop, co-hosted by the National Science Foundation's CISE Directorate and the White House Office of Science and Technology Policy, convened more than 80 researchers from colleges and universities, for-profit companies, nonprofit organizations, and various government organizations. The workshop discussed recent advances in broadband in highly innovative regions of the country, which have agreed to be interconnected with each other via the Global Environment for Network Innovations (GENI) testbed. Workshop participants are expected to use GENI to develop applications in fields such as health, education, energy, economic development, transportation, and public safety. US Ignite will help researchers develop the applications and services and deploy them to end users. One of the main goals of the workshop was to align the applications developed by researchers to cities that could implement them, as well as to understand what resources are needed to launch the services. "The goal here is to take the conversations from [the workshop] and translate them into specific and tangible outcomes and then, as a group, to make sure that we're all held accountable so that we can achieve what we hope will be the next platform for these kinds of innovations," said U.S. chief technology officer Aneesh Chopra.


Home-Computer Users at Risk Due to Use of 'Folk Model' Security
MSU News (05/23/11) Tom Oswald

Most home computer users attempt to secure their machines based on folk models, or their beliefs about hackers or viruses, according to new research from Michigan State University professor Rick Wash. Although people are trying to make the best sense of the situation, they make choices that leave them vulnerable, Wash says. His research has identified eight folk models that people use to decide what security software to use and which advice to follow, ranging from the vague and generic "viruses are bad" to the more specific "hackers are burglars who break into computers for criminal purposes." Wash says the problem of home computer security is compounded because users do not necessarily follow the advice of credible computer security experts. The folk models will help explain why people follow or ignore certain advice. "By better understanding why people choose to ignore certain pieces of advice, we can better craft that advice and technologies to have a greater effect," Wash says. "Security education efforts should focus not only on recommending what actions to take, but also emphasize why those actions are necessary."


Honing Household Helpers
MIT News (05/26/11) Emily Finn

Massachusetts Institute of Technology (MIT) researchers are developing robots that can complete simple human tasks using a hierarchical, progressive algorithm that could reduce the computational power required to perform complex actions. Previous programs that enable robots to function autonomously have been categorized as either task planning or geometric-motion planning, even though an effective robot needs both types of planning to complete tasks. MIT professors Leslie Kaelbling and Tomas Lozano-Perez think the key is to break the computationally complex larger goal into smaller steps, then make a detailed plan for only the first two, allowing the other steps to be dealt with later. "We're introducing a hierarchy and being aggressive about breaking things up into manageable chunks," Lozano-Perez says. The new MIT robots are able to respond to environments that change over time due to their actions, as well as external factors. The robots "do the execution interleaved with the planning," Kaelbling says. The researchers believe this is the best way for robots to complete extremely complex problems. They are planning to build in learning algorithms so that robots will be able to determine which steps should come earlier in the process and which can be postponed.


Single Molecule Performs Multiple Logic Operations Simultaneously
PhysOrg.com (05/23/11) Lisa Zyga

An international team of researchers has developed a single molecule that can perform 13 logic operations, some of them in parallel. The researchers say the molecule, which combines three photochromes to form FG-DTE, is operated by different wavelengths of light and could serve as the basis for molecular computing. "While previous examples of molecular logic systems have been able to carry out one, or a few different logic operations, this molecule can be reconfigured to perform 13 simply by changing the input or output wavelengths," says Arizona State University's Devens Gust. The FG-DTE molecule is the first that can perform certain operations using only two inputs, in this case light with wavelengths of 302 nanometers and 397 nanometers, which allow the molecule to perform addition and subtraction in parallel. "All of these 13 logic operations share the same initial state, that is, the molecule is always 'reset' to one and the same state by the use of green light, irrespective of which logic function is to be performed," says Chalmers University of Technology's Joakim Andreasson. The FG-DTE molecule also can perform non-arithmetic functions, in which the molecule mimics a mechanical rotary switch to connect any one of several inputs to an output, as well as sequential logic functions.


NASA Is Making Hot, Way Cool
NASA News (05/26/11) Lori Keesey

Researchers at the U.S. National Aeronautics and Space Administration (NASA) and the Illinois Institute of Technology collaborated to develop electrohydrodynamic (EHD)-based thermal control, technology that helps remove heat from small spaces, such as in advanced space instruments and microprocessors that tend to overheat. EHD technology uses electric fields, instead of mechanical pumps and other moving parts, to move coolant through tiny ducts inside a thermal cold plate. The waste heat is dumped onto a radiator and spread far from heat-sensitive circuitry. "Any electronic device that generates a lot of heat is going to benefit from this technology," says NASA's Ted Swanson. The system can be scaled to different sizes, including large cold plates, microscale electronic components, and lab-on-a-chip devices, says NASA's Jeffrey Didion. The researchers are currently working with NASA's Timothy Miller to develop EHD pumps in microchannels that are etched onto silicon wafers. EHD technology will be demonstrated in June as part of the Terrier-Improved Orion sounding rocket mission to show that a prototype pump can withstand rocket launch conditions.


Gaming the Archives
Chronicle of Higher Education (05/23/11) Jennifer Howard

Dartmouth College researchers have developed Metadata Games, an experiment that uses crowdsourcing to create archival metadata. The researchers created game interfaces that enable players to tag images, either playing alone or with a human or computerized partner. Lone players come up with tags to describe the images they see, while the dual players try to come up with the same tag or tags. Metadata Games was developed by Dartmouth professor Mary Flanagan, who also directs the Tiltfactor laboratory, which she says is dedicated to exploring "critical play--a method of using games and play to investigate issues and ideas." The researchers tested the game interfaces by selecting about 200 images from Dartmouth's archives, and inviting a small group of players to participate in a pilot project. The players generated 6,250 tags, with about 80 percent of the tags being deemed useful, according to Flanagan. She and Dartmouth archivist Peter Carini stress that archivists should supervise the process, to check the quality of the metadata being compiled, sift out skewed results, and to reassure the archival community that the results will comply with its standards.


Robo-Jeeves Finds and Folds Your Crumpled Shirts
New Scientist (05/24/11) Paul Marks

Routine domestic chores are still a problem for robots because they need to be able to distinguish between distinctive objects, and efforts to tackle such problems were discussed at the recent International Conference on Robotics and Automation. Pieter Abbeel and colleagues at the University of California, Berkeley have devised software that would have Willow Garage's PR2 commercial robot hold up a mystery garment with its two grippers, estimate the length using its twin cameras to detect the lowest point, and record the outline and distinguishing features that may correspond to a collar or button. The software then applies a statistical technique to generate a digital signature that characterizes the item and compares it against a database of common clothes to determine what the robot is holding. Meanwhile, a Cornell University team has programmed a robot to find shoes scattered about a home, using software that contains data on the form and other characteristics of shoes. "These 'little' tasks are in fact very valid challenges that are big problems for machines," says University of Hertfordshire roboticist Kerstin Dautenhahn.


Tracking Whale Sharks With Astronomical Algorithms
Wired News (05/24/11) Brandon Keim

Researchers at the ECOCEAN Whale Shark Photo-Identification Library are using algorithms designed to guide the Hubble telescope to identify the world's biggest fish by their spots. "It lets you recognize and track animals without marking them, and it's permanent," says ECOCEAN researcher Al Dove. The project started with Princeton University's Ed Groth developing algorithms to compare photographs of the night sky and determine what star patterns they had in common. Whale shark enthusiast Jason Holmberg adapted the algorithms to analyze whale sharks' spots. That work evolved into the ECOCEAN library, which contains 32,000 photographs of about 2,800 different whale sharks. The researchers use the data to keep track of whale shark population trends. The algorithms also have been customized for use on the whisker patterns of polar bears and the fin shape of humpback whales.


Hands-Off Training: Google's Self-Driving Car Holds Tantalizing Promise, but Major Roadblocks Remain
Scientific American (05/23/11) Nick Chambers

Google is lobbying Nevada to be the first state to allow self-driving cars to be legally operated on public roads. Google's robot car fleet, which consists of six Priuses and one Audi, has driven more than 240,000 kilometers with minimal human intervention and just one accident. The robot cars are equipped with off-the-shelf components, including two forward-facing video cameras, a 360-degree laser range finder, four radar sensors, and advanced global positioning system units. The cars only work if the route is driven ahead of time by a human driver so that one of the test cars can map the area. The mapping data is stored at a Google data center and a portion of it is uploaded into the car's hard drive before a trip. "Computers are famously devoid of common sense, and you can think of this pre-mapping as a way to bootstrap some common sense into the car," says Google's Chris Urmson. However, Massachusetts Institute of Technology professor John Leonard thinks that major technological barriers in robot perception need to be overcome before self-driving cars can be used on a large scale. He has been focusing on Simultaneous Localization and Mapping technology, which would allow a vehicle to drive and map its surroundings at the same time, a method that would greatly increase the self-driving car's ability to cope with constantly changing information.


Abstract News © Copyright 2011 INFORMATION, INC.
Powered by Information, Inc.


To submit feedback about ACM TechNews, contact: [email protected]
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.
Non-Members: Unsubscribe