Google

Friday, December 21, 2007

IBM Gets Into Semantics With New E-Mail Search Tool

IBM's new OmniFind Personal E-Mail Search tool is designed to help heavy e-mail users sift through the glut of information residing in their in-boxes. The technology works by intelligently matching a query against predefined concepts, such as persons, phone numbers, addresses, meetings, presentations, documents or schedules and relationships amongst these concepts, IBM said.

Heavy e-mail users sometimes have trouble finding the needle of information they need in the haystacks of old missives lingering in their in-boxes. That's the sort of user IBM (NYSE: IBM) has targeted with its new OmniFind Personal E-Mail Search (IOPES) tool.

Rolled out Thursday, IOPES is a free, semantics-based search tool designed to make searching the "vast personal database" that is most people's inbox easier by identifying the most relevant information in a search query and extrapolating what the user is attempting to find, according to IBM.

"With gigabytes of e-mail storage readily available to nearly everyone, e-mail has evolved from a simple communication tool into a personal database where we retain vast amounts of valuable information," said Douglas Wilson, distinguished engineer and chief technology officer of Lotus.

IOPES for Microsoft (Nasdaq: MSFT) Outlook and Lotus Notes is available at no charge through IBM's AlphaWorks site.

Search Grid

Developed jointly by researchers at IBM labs in California, Israel and India, IOPES offers more refined search capabilities than a simple keyword search. Instead, the e-mail search tool uses semantic technology that can interpret incomplete queries and make associations based on the underlying meaning of words frequently used in corporate e-mails.

The software was created using IBM's Unstructured Information Management Architecture (UIMA), an open source software framework that helps organizations build new analysis technologies. Companies are able to realize more value from their unstructured information by discovering relationships, identifying patterns and predicting outcomes, according to IBM.

The technology works, IBM said, by intelligently matching a query against predefined concepts -- such as persons, phone numbers, addresses, meetings, presentations, documents or schedules -- and relationships amongst these concepts -- such as a person's phone number or address.

To accomplish this, IBM researchers compiled an index of keywords typically found in business e-mails and then created a list of concepts and relationships among those terms. That data was built into IOPES so that when a user enters a query -- for example for "Bob address" -- the application first runs through its index of terms and relationships to return the most accurate results.

Users can further refine IOPES' search functionality by adding their own user-defined concepts using the tagger utility that can be shared between individuals and used to create a more personalized search system, according to IBM.

Semantically Speaking

Creating better search tools based on semantics has been a widely talked-about goal for online search engines. However, less work has been done in the areas of desktops, e-mail and other data searches, leaving users to struggle with keyword searches.

"This is a good thing. It's good to see that someone is bringing this into the e-mail or desktop search category," said Kenneth Poore, a Forrester Research analyst. "The industry in general has been kind of underserved by the keyword-based search tools that Microsoft has with their Windows Desktop Search, and of course Google has their desktop search."

The problem with the current generation of desktop and e-mail searches is that they are "so literal and so keyword-based" that they are "not hitting the mark, and a lot of users are coming away a little disenchanted with the dearth of capability" in those products, Poore told TechNewsWorld.

"They want more of 'do what I mean' or 'find what I want' kind of searches rather than find exactly what I type in," he continued. "IBM has done a lot of work with semantic analysis. Doing semantic search for e-mail really opens that up. Being able to go out and do more conceptual searching instead of literal searching adds a lot of value."


Microsoft Cagey on Details as It Sets IE 8 Time Frame

Microsoft started the countdown to the next iteration of its Internet Explorer Web browser, IE 8; the company is promising a beta version in early 2008. Typically, Microsoft isn't very shy about revealing a few details about upcoming features in such offerings, but its latest announcement had little to say about the new functions to be found in the beta.

The next iteration of Microsoft's (Nasdaq: MSFT) Web browser, Internet Explorer 8 (IE 8), will roll out in beta form early next year, the company said Wednesday. The long-awaited news comes just days after the earlier-than-expected launch of the second beta version of Mozilla's Firefox 3 Web browser.

The eighth version of the browser won't be hamstrung with the difficulties IE 7 had following its release in October 2006, Microsoft said, due to improvements Microsoft made to IE 7's Cascading Style Sheets (CSS). The browser passed a significant Web standards milestone, and as of last week, IE 8 now "renders the 'Acid2 Face' correctly in IE 8 standards mode," said Dean Hachamovitch, IE Team general manager.

"Acid2 is one test of how modern browsers work with some specific features across several different Web standards," he continued. "With respect to standards and interoperability, our goal in developing Internet Explorer 8 is to support the right set of standards with excellent implementations and do so without breaking the existing Web. This second goal refers to the lessons we learned during IE 7."
Web Standards Focused

With Internet Explorer 8, Microsoft is making a concerted effort on interoperability, ensuring the browser supports certain accepted Web standards. It's something the software maker did not focus on until it began working on IE 7 in 2005. Before then, Web developers designed their applications with an eye on compatibility with the latest version of Internet Explorer rather than the jumble of industry, de facto and open standards.

The rise in popularity of Web standards such as CSS and Really Simple Syndication (RSS) and growing competition from open source browsers like Firefox and Opera forced Microsoft to rethink its position.

"The key goal ... is interoperability. As a developer, I'd prefer to not have to write the same site multiple times for different browsers. Standards are a (critical) means to this end, and we focus on the standards that will help actual, real-world interoperability the most," Hachamovitch said.

"This is a big push towards standards compliance," Rob Enderle, principal analyst at The Enderle Group, told TechNewsWorld. "IE is the product that defines how Web sites are built, so its move to standards should be, and initially has been, very well received."

Passing the Acid2 test developed by the Web Standards Project, an advocacy group, is a significant achievement. Firefox and Opera failed this test in the past, and it is believed to be the most comprehensive third-party test for broad standards in the market, Enderle pointed out.
Browsing Through IE 8

Though the final version of IE 8 will be a full release, Microsoft has remained relatively quiet about what new improvements and features will be included in the final form.

Microsoft trumpets new features early and often, "but they appear to be keeping some of the features under wraps," Enderle noted.

Judging by Hachamovitch's remarks, Microsoft's unusual silence is intentional.

"For IE 8, we want to communicate facts, not aspirations. We're posting this information now because we have real working code checked in and we're confident about delivering it in the final product."

With the software maker keeping mum about what it plans for IE 8, one good sign is that Microsoft is not influenced as much by what its competitors Firefox, Opera and Safari are doing, Enderle said.

"Microsoft actually appears to be listening to customers and developers on this cycle, which normally bodes well for the product. Focusing on competitors too much often results in a product that lags significantly behind what the others have to offer. And Microsoft needs to lead with IE, not follow," he explained.

"I'd expect a big push on media, major security enhancements, and some solid usability work with the final release," Enderle added.
Tread Carefully

Though he cautioned Microsoft against focusing too much on what its competitors are doing, Enderle suggested the software maker take a look at the mobile version of Apple's (Nasdaq: AAPL) Safari browser. "The big advantage Safari has is on the iPhone, and addressing this shortcoming in the mobile version of IE is critical because it's crippling their phone effort right now."
Whatever Microsoft decides to do with IE 8, it is imperative they get it right, Enderle stated.

It may be one of the major things that define Microsoft in the last part of the decade, as the code is tied to almost every platform they have in some way or other. Getting this right or wrong could have a major impact on Microsoft's future," he concluded.


Trojan Pulls a Fast One With Google Text Ads

End users who click on seemingly legitimate Google ads may be at risk of infection by a Trojan that substitutes rogue ads for the real thing. Google and the companies that pay for genuine ads are also victimized, because the pretenders usurp traffic and potential revenue.

Malware is replacing Google (Nasdaq: GOOG) text ads with ads from another source, according to BitDefender. The virus, Trojan.Qhost.WU, is using the host's file to redirect the initial query sent to the Google Adsense servers to a malicious host, according to an advisory issued by the firm.

The host's file is the first step in the name/IP (Internet protocol) translation process; if an entry is located in this file, the domain name server is not queried. By supplying a false entry, the malware is able to redirect queries to a rogue server.

Who's at Risk



End users who click on the seemingly legitimate ads are at risk, as they likely carry additional malware. Google and the companies that pay for genuine ads are also victimized, because the pretenders usurp traffic and potential revenue.

To see if a computer has been infected with this virus, BitDefender advises users to investigate whether the host's file is providing local storage for domain name/IP mappings that contains a line redirecting the host to page2.googlesyndication.com.

From the command line or from Start-->Run, issue the following command: ping -t pagead2.googlesyndication.com. The response should look similar to this, according to BitDefender: Pinging pagead.l.google.com [6x.xxx.xxx.xxx] with 32 bytes of data, where the x's represent digits.

"If you are not infected, the first digit will be a 6 (as in the example). If you are infected, the first digit will be a 9," said BitDefender.

Trojan.Qhost.WU is not spreading fast and poses a "medium" risk of damage, according to the advisory.

Not Unusual

While the target may be a little different, this particular Trojan is just another variation of typical phishing malware, Dmitri Alperovitch, principal research scientist with Secure Computing, told TechNewsWorld.

"We have been seeing attacks like this for the last two to three years, where the virus changes the internal setting to point the user to a different server," he said.

At their core, all of these hack attacks intercept a resolution from the browsers to the DNS (domain name system) server via a simple modification to the Windows system file, he explained. "No query is made to the real DNS server."

A more dangerous variant is the Zlob virus, which infects users by masquerading as a video compression algorithm necessary to view a particular video.

The malware that is subsequently downloaded replaces resolutions not for just one domain name, but for an entire configuration of DNS servers under the control of a malicious group.


S5 Wireless: Going Where No GPS Has Gone Before

A GPS device may be able to tell you which roads to take to drive to the southernmost tip of Chile, but take it inside any building and odds are it won't know where you are. A new company, S5 Wireless, wants to create an alternative tracking system that works well indoors. S5's system, however, works differently than GPS, meaning that if it takes off, it would be more a complement than a replacement.

In one high-tech thriller after another, the hero attaches a tiny tracking device on the villain and follows him as blinking dot on a computer screen. In real life, this kind of technology would be great for tracking pets or kids, even packages or luggage -- anything that tends to wander. However, it doesn't really exist.

There are GPS devices, of course, but strap a half-pound GPS collar to a dog and you'll realize it's far from "Mission Impossible." GPS-enabled cell phones are becoming more common, but they have problems, like accuracy indoors, and they aren't cheap.

A Utah company, S5 Wireless, is looking to bring reality closer to the movies, with small, cheap chips that can be powered by a single battery for up two years and tracked indoors and outside, over long distances.

For instance, an S5 chip could go into a dog collar, complete with a battery, in a package about the size of a stick of gum that costs $3 to $4 to make. When the battery runs down, it's time to buy a new collar.

The same concept could be applied to a kid's backpack, with an antenna running through the strap.

"It's like a poor man's LoJack or OnStar," said David Carter, S5's chief executive.

Triangulating Signals



The drawback to the technology is that unlike the Global Positioning System, which is quite literally global, S5's technology would only work where the company has a network of stations to receive S5 signals. S5 is planning to start building those in some major U.S. cities next year.

What the chips do is basically GPS in reverse. GPS satellites operated by the Air Force send signals to receivers in devices like car navigation systems. Those receivers need a line of sight to the sky, so they work only outdoors, and are fairly power-hungry.

By contrast, the S5 chips send radio signals that will be picked up by receivers S5 plans to build. By measuring the difference in the time the signal takes to arrive at three different receivers, S5 can compute the location of the sender to within about 30 feet outdoors, or 45 feet indoors. The same principle can be used for locating some cell phones in an emergency but is much less accurate. LoJack equipment also sends radio signals from a stolen car, but it uses a lot of power and is expensive compared to S5's chips.

GPS: Not Entirely Housebroken

"The thing that struck me the most is that their solution allows for very good indoor coverage," said Belgium-based analyst Dominique Bonte, who covers GPS technologies for ABI Research.

Bonte said the weakness of GPS indoors vexes those who want to provide location-based services, like cell-phone messages advertising sales at nearby stores.

Companies like Intel (Nasdaq: INTC) have experimented with using WiFi hotspots in known locations to fill gaps in GPS. Some GPS devices incorporate motion sensors that estimate the user's location based on movement after the GPS signal is lost.

"There's a lot of activity, a lot of companies, and of course a lot of venture capital which is being made available for those kinds of initiatives," Bonte said. "There is a big belief that whoever comes up with something that's cheap, that works and is available will hit the jackpot there."

Giving It Away

To gain a foothold, S5 will give away the designs for its chips, letting anyone make their own or incorporate the functions into existing chips, like those in cell phones. It plans to make money by charging for the location service, though at low rates, around US$1 a month, Carter said.

S5's technology isn't ideal for navigation devices, since the chip doesn't know where it is. It couldn't, on its own, plot its position on a map the way a car navigation system does. Instead, it is S5's data center that knows where the chip is.

If you're tracking a dog -- or a villain -- that's not an issue. The dog doesn't need to know where it is, but the S5 network can tell you via Web browser or cell phone where it is.

Apart from pets and kids, Carter envisions the system being used to track valuable equipment on construction sites and in hospitals, much like LoJack protects cars.

If you don't trust FedEx and UPS to track your package, you could slip a tracking device into it before you ship it, and you'll know where it is. The company has also received grants from the Department of Homeland Security to study the use of its chips in tracking shipping containers. S5 chips could even be built into cell phones to supplement GPS chips where reception is weak, like indoors.

As a bonus, S5 chips could transmit small amounts of data generated by other devices. For instance, a diabetes patient's glucose meter could be monitored remotely.

Major City Rollout

S5 plans to piggyback on existing cell-phone towers and antennas in building out its network, though officials would not say how much they expect the build-out to cost. To pinpoint a chip's location, S5 needs three receivers within the signal's range, about a mile in cities, Carter said. The company plans to cover "several" major cities next year and 35 cities within three years.

That's a tall order, but the venture's main backer lends it credibility. Billionaire Craig McCaw, who founded one of the country's first cell-phone companies, is S5's majority investor. He also backs wireless broadband provider Clearwire.

S5 hasn't announced any build-out partners yet.

Remarkably, S5 plans to use free, unlicensed spectrum in the 900 megahertz band, which is already crowded by cordless phones. Steve Chacko, S5's director of product marketing, likened the feat of picking up those signals from miles away to extracting a needle from a haystack. However, he said sophisticated low-power radio technology makes S5's plan viable.


Intel Asks Devs to Help Get LANs, SANs to Play Nice


FCoE is a proposed specification that will allow Fibre Channel storage area network traffic to run over Ethernet. By consolidating LAN and SAN traffic onto a single fabric, FCoE aims to simplify network infrastructure in the data center and give organizations more flexible options for deploying and managing existing SANs.

Intel (Nasdaq: INTC) has released a software initiator package designed to promote the development of Fibre Channel over Ethernet (FCoE) solutions for Linux, the company announced Tuesday.

Now available for download, the software is licensed under the GNU General Public License version 2 and is open to all FCoE developers for source code download, code contribution and feedback. It includes a target simulator so Linux developers can test and modify the FCoE software stack, Intel said.

"Server virtualization, database growth and compliance requirements have created a greater need for network storage," said Pat Gelsinger, senior vice president and general manger of Intel's Digital Enterprise Group.

"FCoE makes it easier and less expensive to connect servers to the SAN," Gelsinger added. "We expect the open source community to use this initiator stack to create reliable, standards-based FCoE solutions for their customers."
Proposed Standard

FCoE is a proposed specification that will allow Fibre Channel storage area network (SAN) traffic to run over Ethernet. By consolidating LAN (local area network) and SAN traffic onto a single fabric, FCoE aims to simplify network infrastructure in the data center and give organizations more flexible options for deploying and managing existing SANs.

Vendors backing the standard, which was proposed in April to the T11 Committee of the American National Standards Institute (ANSI), include Brocade, Cisco (Nasdaq: CSCO) , EMC (NYSE: EMC) , Emulex, IBM (NYSE: IBM) , Nuova, QLogic and Sun Microsystems (Nasdaq: JAVA) , as well as Intel.

Intel is a member of the T11 Fibre Channel standards body.
'Key Capability'

"Fibre Channel over Ethernet will be a key capability for our customers offering seamless server and storage access in the data center," said Jayshree Ullal, senior vice president of the data center, switching and services group at Cisco Systems.

"The emergence of 10 gigabit Ethernet bandwidth combined with Cisco's proposed extensions to Ethernet, enables a lossless and resilient fabric for data center I/O (input/output) consolidation," Ullal added. "Cisco is pleased to see Intel taking a leadership role in FCoE."

The new FCoE initiator code is based on a specification being developed by the T11 in the FC-BB-5 work group. The FCoE specification is expected to be completed in 2008.
'A Critical Step'

"The classic challenge for any storage networking technology, whether Fibre Channel, FCoE or iSCSI, is developing support for the most important operating environments," Richard L. Villars, vice president of storage systems for IDC, told LinuxInsider.

Fibre Channel, because of when it came out and where it's used, is the dominant technology for all the Unix platforms and the mainframe world, as well as growing in importance for the Windows, VM (virtual machine) and Linux worlds, Villars added.

"FCoE is a critical step for companies that want to leverage their past investments in Fibre Channel SANs and networking software," he explained. "This makes the transition to FCoE smoother as companies mix the older operating systems with the newer ones where much of the application development is going on today."












No comments: