Wednesday, December 27, 2006

Samsung demonstrates fuel cell laptop

According to Akihabaranews, Samsung is now offering a high-density fuel cell storage system for its Q35 laptop. The power system is contained in a large dock that the laptop sits on top of.

Samsung claims that the fuel cell offers an energy density of 650Wh/L, which is about four times as much as competing offerings. The total energy storage is an impressive 12,000Wh which, depending on the laptop's power settings and usage could theoretically power the laptop continuously for a whole month without the need for recharging.

One of the big breakthroughs Samsung is touting for their new fuel cell system is a major reduction in noise level: the new unit is said to be no louder than conventional laptops. While the docking station is somewhat bulky, it still qualifies as being portable, which is a claim that other fuel-cell solutions have not always lived up to.

Besides the advantage of much longer battery life, fuel cells have other advantages, such as near-instantaneous recharging time by simply adding more fuel. Companies such as Casio were demonstrating laptop fuel cell prototypes as early as 2003. Some even predicted that devices as small as cell phones could use fuel cell batteries, but the technology has taken some time to become ready for consumer use. As fuel cells become smaller, lighter, and more powerful, they may eventually take over from conventional batteries for all kinds of applications. And the days of mobile computing users rushing to find power outlets may become a thing of the past.

Science's Breakthrough of the Year

As the end of the year approaches many publications are releasing their top 10 lists for the year and Science is no exception. Last year Science named evolution as its top breakthrough of the year, but was accused of pandering to the political/religious debates that were/are raging throughout the world, especially in the United States. This year, Science (open access) named a breakthrough that has no connections to politics or religion: the proof of the Poincaré Conjecture by Russian mathematician Grisha Perelman.

The Poincaré Conjecture was originally proposed by Henri Poincaré in 1904 and deals with the topology of everyday objects, namely what, in topological terms, defines a sphere. The Conjecture remained unsolved for almost 100 years, although not for lack of trying, and in the year 2000 the Clay Mathematics Institute (CMI) named the Poincaré Conjecture as one of its six Millennium Problems. These problems have solutions that have eluded mathematicians for years and carry a US $1,000,000 prize to anyone who solves them (either in a positive or negative manner). As stated in the CMI's official problem declaration, the Poincaré Conjecture asks

If a compact three-dimensional manifold M3 has the property that every simple closed curve within the manifold can be deformed continuously to a point, does it follow that M3 is homeomorphic to the sphere S3?
Clear, no? Well for those of us who do not hold advanced degrees in topology or geometry, it asks a fairly simple question, namely "Can an arbitrary closed surface be turned into a sphere by only stretching it appropriately?" This question has formed a cornerstone in the field of mathematics known as topology, which studies the geometry of surfaces undergoing deformations. Poincaré suggested that ALL three-dimensional surfaces that had no holes could be turned into a three-dimensional sphere without needing to tear a section apart or stitch two sections together. He also suggested the converse to be true, that a surface with a hole could NEVER be turned into a sphere without tearing or sewing sections. A real world example from breakfast would be that a muffin can never be turned into a doughnut: the hole in the doughnut cannot be created in the muffin by simply stretching it—one would need to tear a hole in the muffin to make it ever resemble the doughnut.

Over the years, special solutions for specific dimensions were developed. For one and two dimensions, the proof was trivial; seven or more dimensions were handled by a proof from Stephen Smale, developed in 1960. Smale then extended his proof to all dimensions greater than or equal to five, for which he was awarded the Fields Medal in 1966. More than 20 years later, Micheal Friedman proved the Conjecture for the four dimensional case in 1980 and was awarded the Fields Medal in 1986. Poincaré had now been proven for ALL dimensions EXCEPT three, the original dimension for which Henri Poincaré first proposed the problem.

The final unproven dimension fell in a series of papers published to the web by Grisha Perelman, all available to the curious for free from arXiv.org. Dr. Perelman extended previous work done by Richard Hamilton, a mathematician who proposed that an arbitrarily lumpy space could "flow" towards a smooth space—imagine a Koosh ball morphing into a soccer ball—through equations akin to the heat equation, and named the process "Ricci flow." In Ricci flow, lumpy areas and areas that are highly curved tend to smooth themselves out until in entire surface has a constant curvature. In the absence of any major problems, this idea of Ricci flow could show how an arbitrary surface could be morphed by simple stretching into a smooth sphere, thus providing a method to prove or disprove Poincaré's Conjecture. However, problems arose; singularities, such as necks (thin areas between two larger areas, e.g. the bar in a dumbbell weight) would pinch and close themselves off, forming two separate objects with a uniform geometry, violating the rules of topological stretching. Perelman had to work for many years to find a way to overcome this problem. In November of 2002, he published his first paper on the subject with a new quantity added into the mix, what he called Ricci flow "entropy", borrowing from a term in statistical mechanics that tends to increase until a equilibrium is reached. This entropy idea proved that the problems in Hamilton's work—the singularities—could be overcome, yet Perelman still faced other road block to overcome before a full proof could be developed. In his subsequent articles, he showed that these problems areas would occur one at a time, as opposed to all at once. He then went on to show that it could be "pruned" through surgery before it would cause a problem with the Ricci Flow.

This series of papers was published in 2002 and 2003, yet it took the mathematics community at large three more years to accept this solution as a true proof to the Poincaré Conjecture. However, the story does not end there. In 2006 the International Mathematical Union announced that it had awarded the Fields Medal to Grisha Perelman, but he declined it. In a rare interview with The New Yorker, Dr. Perelman announced that he was retiring from mathematics, stating he was disheartened by what he viewed as ethical lapses by some of his colleagues. This New Yorker article has caused quite a commotion in the mathematical world, with people claiming their words were distorted and threats of lawsuits traded. This has caused a black cloud to hang over what is the greatest mathematical breakthrough of the millennium (so far). Fortunately mathematical proofs are not affected by the feelings of those who create them. The proof of the Poincaré Conjecture remains atop the list of the greatest scientific breakthroughs of the year.

Thursday, December 21, 2006

Apple Bug-Hunt Begins

Apple Computer will soon be a member of the "month of bugs" club.

On January 1, two security researchers will begin publishing details of a flood of security vulnerabilities in Apple's products. Their plan is to disclose one bug per day for the entire month, they say.

The project is being launched by an independent security researcher, Kevin Finisterre, and a hacker known as LMH, who declined to reveal his identity.

Some of the bugs "might represent a significant risk," LMH said in an e-mail interview. "Others have a lower impact on security. We are trying to develop working exploits for every issue we find."

The two hackers plan to disclose bugs in the Mac OS X kernel as well as in software such as Safari, iTunes, iPhoto, and QuickTime, LMH said. Some of the bugs will also affect versions of Apple's software designed to run on Microsoft Windows, he added.

LMH was one of the brains behind the recent Month of Kernel Bugs project, which exposed flaws at the core of several different operating systems.

Wednesday, December 20, 2006

Polar Rose wants to recognize faces in web photos

Swedish startup Polar Rose today took the wraps off of their plan to make faces in photos searchable. Searching within images has always been tricky for machines, but Polar Rose hopes that a combination of mathematics and user intelligence will help to build an accurate recognition engine.

The software works by constructing 3D models of the 2D faces found in photographs, a technique that gives a "significant boost to recognition," according to the company. The 3D model makes it easier to construct a unique facial "fingerprint," but does little to match that fingerprint with an individual. That's where users come in.

Polar Rose hopes to make the system match up the faces with names by providing a browser plug-in for Firefox and Internet Explorer. When the plug-in locates people (or what it thinks are people) in online photos, it puts the company's rose logo in a corner of the picture. Clicking it will allow web surfers to tag photos with names and other information. Once the system is trained to recognize a particular person, it can locate that same person in other photos, even if the new photos contain no metadata and are taken from a different angle. All tagging information is kept by Polar Rose, which means that all users get the benefit of everyone's tags.

The goal is to provide a workable search engine for people in photos, but the company is also opening up the technology by releasing a set of APIs early next year. The royalty-free APIs will allow for the creation of new applications that go beyond anything dreamed up during snowy nights in Malmö, where Polar Rose is based.

Thursday, December 07, 2006

Open XML becomes a standard

Microsoft Corp. has won approval for its Offfice Open XML document format from international standards body ECMA International.

ECMA's General Assembly voted by 20-1 in favor of the standard at a meeting in Zurich on Thursday afternoon, and will now submit the format to the International Organization for Standardization (ISO) for its approval. The vote against came from a representative of IBM Corp.

A standardized document format will make it easier for competing software companies to develop products that can interoperate with one another and to edit the same documents. Products meeting the standard could find favor with governments, or other organizations concerned about interoperability.

Interoperability is vital for the preservation of archive information, according to Adam Farquhar, head of e-architecture at the British Library and a member of the committee that worked on the standard at ECMA. The British Library archives electronic documents, but must deal with whatever format they arrive in. The development of interoperable software tools will make that work easier.

To help developers digest the standard, the committee has published a 14-page white paper explaining it.

That accessibility is important if Microsoft is to win developers over to its document format, as the company faces competition in the standards industry just as it does in the software market. A rival document format, OpenDocument Format (ODF), has already won approval from ISO, and was published as an ISO standard last month. ODF is used by office productivity suites such as OpenOffice.org or Sun Microsystems Inc.'s StarOffice, and has gained the support of other companies, including IBM. Government officials in France and the U.S. commonwealth of Massachusetts have recommended adoption of ODF as a government standard.

Monday, December 04, 2006

Open Document Format published as ISO standard

The International Organization for Standardization (ISO) finally published the Open Document Format (ODF) as an official standard last week after approving it as an international standard last May. The ODF file format—the XML-based open format for text, spreadsheet, database, and presentation files—is now published under the standard name of ISO/IEC 26300:2006.

ODF is being championed as the accessible file format by accessibility groups, as the open standard must face public peer review to evaluate its accessibility as opposed to other, proprietary formats (such as .doc). Since it is specified as a public document, the idea is that anyone can create software for it under any platform at any time for those with disabilities without worrying about royalties or implementation barriers, not to mention those who simply like being able to move files around without concerning themselves with incompatible formats.

This move will likely give an advantage to ODF over other open standards, such as the Office XML standard that Microsoft has been working on. There are already a number of companies and organizations worldwide that support the use of ODF, and various software, such as OpenOffice.org, already generates ODF files as the native file format. Additionally, Google's web-based office apps support ODF, and even Microsoft has conceded that it will support a plugin for Microsoft Office that will convert files to ODF.

Zudeo: a high-def version of YouTube

Another day, another peer-to-peer file-swapping firm trying to go legit. This time it's Azureus, makers of one of the most popular BitTorrent clients on the Internet. The company has just launched its own video sharing site, dubbed Zudeo, which hopes to stand out from a crowded field of contenders by offering high-definition content and storage space for massive files.

The site, still in beta, allows the usual mix of tagging, uploading, and free viewing, though content is quite limited and features are still sparse. Users first need to download the Azureus application to get content, though, adding a layer of complexity to the normal process.

In return for using a BitTorrent client, Zudeo offers full HD-quality videos and gigabytes of storage space. It's free to content producers, who need to use the Azureus Director Edition to upload their creations.

Azureus isn't the only company using BitTorrent to distribute massive video clips; there's also, err, BitTorrent. BitTorrent (the company) has its own plans to become a content distributor, and it just signed deals with 20th Century Fox, Lionsgate, MTV Networks, Paramount, and Starz that will allow it to offer their content for download in the new year (in addition to deals it signed earlier in the year). The company has also closed another $20 million funding round, a sign that investors still have confidence in peer-to-peer distribution models (or that they desperately want to create a YouTube-style success story).

Hacker boots out evolution archive from Google

Webmasters aren't pleased when they run into problems with Google, and Wesley Elsberry was no exception. Elsberry, the webmaster of the Talk.Origins archive, made his displeasure public in a blog post after Google booted his site from their search index. This sort of thing happens frequently, but a public response from Google is uncommon, and it sheds some light on how Google handles its relationships with webmasters. It's also a great example of how far you'll get with a problem if you call Google on the phone.

Talk.Origins is devoted to discussions of evolution and the perceived shortcomings of the Intelligent Design and creationist movements. The group behind Talk.Origins runs several websites, one of which was hacked on November 18. The hacker added a collection of invisible spam links to certain webpages, links that dealt mostly with the less-pleasant sides of human sexual experience. Google's algorithms noticed the spammy links within a few days and delisted the site from the Google Index for 60 days.

Elsberry was angry that he wasn't notified in advance of this decision, and claimed that Google even refused to tell him what the problem was. "So, what, precisely, was causing Google to not like us anymore?" he writes. "The essential lesson here is that Google would not tell us. That isn't mere caprice; that is Google policy." Elsberry's attempts to get access to a Real Live Human™ on the phone proved fruitless as well, but he reports that he luckily found and fixed the problem quickly.

After the story was posted on Slashdot, Google's Matt Cutts investigated the issue and wrote a blog post of his own, outlining the steps that Google had taken. By November 27th, the site was listed as "hacked and spammy" and was flagged as "penalized." The next day, the Google folks emailed several addresses at talkorigins.org about the problem, and included specific links to the problem pages. Elsberry never got the messages.

But Cutts argues that it's ultimately the responsibility of webmasters to safeguard their own domains, and that Google can't provide too much information about its practices or the types of problems it flags (this could make it easier for spammers to evade automated scrutiny). He agrees that Google could do a better job of getting the word out, but concludes, "Please give Google a little bit of credit, because I do think we're doing more to alert webmasters to issues than any other search engine."

Being indexed by Google is all but a prerequisite for running a successful website, which continually puts Google in the position of having to play the roles of judge, jury, and executioner. While they have no control over a site's bandwidth, one decision by Google can effectively put many sites out of business. Talk.Origins won't be one of them, though; the site has already been reinstated and should soon be fully accessible through the search engine.