Bienvenue dans le troisième âge de la recherche

September 10th, 2008 by Guest Author
Posted in Global, Guest Authors | No Comments »

Welcome to the Third Age of Search
By Nicolas Kayser-Bril  Translated by Mark Cramer

At the dawn of time, while France loved the Twingo as much as she hated Aimé Jacquet, search engines were still called AltaVista and Lycos. In response to a query, they searched their index, searching for words in the requested pages in memory. They proudly returned with approximate results.

Then came 2000, the third album by Gerald de Palmas and the success of the revolutionary engine Google. In contrast to its predecessors, Google takes into account the reputation of various sites to establish its results. PageRank, the obsession of all webmasters, reflects the relative popularity of each page. To be at the top of Internet search, the page’s content is no longer enough, we must take into account its relations.

Eight years later, the day of an ordinary marriage, Google entered the third age. The results will henceforth be organized based on previous searches, announced the mastodon. Although the function existed previously in Beta version, it now applies to all those who have a Google account, such as AdSense, iGoogle and GMail, among others.

Once you are identified, Google records your history and uses it to provide more relevant results. You can always disable it by clicking “History,” top right on the results page, and then “Pause”.

Two years ago, AOL mistakenly published the search histories of 650,000 subscribers. Although anonymous, data made it possible to retrace the private lives of multiple users.

Surf Canyon: Sharpener of Searches

Nevertheless, the personalization of search engines attracts even the startups. Surf Canyon, a complementary module for Firefox and Internet Explorer, observes the links clicked on the results page of Google and then deduces the subject of your research.

If you ask Google “haski” for example, you will be presented with pretty photos of canines. Few pages, on the other hand, of your preferred media magnate. Find an appropriate result, click on the icon shaped target and Surf Canyon will find other relevant pages, buried in the mass of subsequent results. Somewhat like Google’s “similar pages,” but more useful.

Alain Joannès spoke of such last November, seeing a way to refine searches without giving away too much control. In a new post, he describes the benefits of the program in action. Time saved: ten minutes for a search. Following his pattern, we understand how Surf Canyon can refine a search, digging in the vein the most relevant.

 

For experienced users, such technology may seem superfluous. However, most complaints from Internet users remain very sketchy.By analyzing more than a billion searches, four scientists found that more than half of them use not more than two words and only one in five has a logical operator – specifications that enable clarifying searches (if that seems obtuse, Google has a very good tutorial).

To improve productivity on the web, it would be better to train users. Less dangerous and probably more effective than an Orwellian inventory past searches.

Stealth sighting – Image Search Engine GazoPa

September 10th, 2008 by Charles S. Knight
Posted in Global, Newcomers, News, Verticals | No Comments »


GazoPa is a similar image search service on the web in private beta by Hitachi. Users can search images from the web based on user’s own photo, drawings, images found on the web and keywords.

GazoPa enables users to search for a similar image from characteristics such as a color or a shape extracted from an image itself.

There are abundant quantities of images on the web, however many of these simply cannot be described by keywords.

Since GazoPa uses image features to search other similar images, a vast range of images can be retrieved from the web. GazoPa is a new visual search service that can navigate users to new territories on the web.  Source: GazoPa

Search!o the tabbed search engine from Search.io

September 10th, 2008 by Charles S. Knight
Posted in News | No Comments »

Search.io is a cool and innovative website which aims to simplify your searching needs by bringing a lot of sites in a single page.

It allows you search the popular sites which fall in a particular category. The categories include search engines for music, images, books, blogs, news, fonts, recipes, jobs, torrents, tutorials people and many more.

It has a simple user interface and is also very fast in delivering results. As you hover your mouse pointer on the selection box, it immediately pops up the list of categories to choose from.

As you type select a category and enter a search query, it shows the relevant websites in tabs on a single page. You can navigate between the tabs by clicking them or by keyboard shortcut- Alt + ‘ Tab Number ‘ . As you can see, I pulled up the ReadWriteWeb site under the Search Web 2.0 Sites category. You can also submit results to various social media sites from search.io.

Source: makeusof.com

Cluuz provides consumers with search clues!

September 10th, 2008 by Charles S. Knight
Posted in News | 1 Comment »

Here is a timely review of alternative search engine Cluuz.  Sprylogics is the owner of the Cluuz search platform.  Developer of next generation semantic search technologies and intelligence-based risk mitigation software and services, Sprylogics, has unveiled a new challenge to the way that users find information online.

While tech heavyweights Microsoft Corp. and Yahoo Inc. spend billions trying to wrestle control of the web search industry from Google Inc., a growing collection of niche search engines are rethinking the way users find information – and they are catching the attention of the big three. By developing tools that scour web pages for the actual meanings of words, rather than merely providing a list of sites with matching keywords, these challengers hope to create services that help users find what they’re looking for faster. It doesn’t mean these upstarts are trying to beat Google at its own game.

That’s not the point, says Michael Frank, chief executive officer of Toronto’s Sprylogics International Inc., the makers of Cluuz.com, a semantic search engine attempting to get users to rethink the way they find information online. Mr. Frank surmises, ”What consumers want is to be able to find information faster, and they want clues to help them find their way to that information faster. What we’re doing is quite unique and nobody can do what we’re doing.”

According to Mr. Frank, search engines such as Cluuz use the science of semantics – the study of meaning in language – to produce more relevant searches.While the site works as a stand-alone search engine, it could also work if it were rolled into existing offerings at Google, Yahoo or Microsoft’s MSN.

In July, Yahoo announced it was opening up its search index data to developers as part of a project it dubbed Build your Own Search Service (BOSS). In a press release announcing the BOSS program, Cluuz was one of four services Yahoo cited as examples of innovative search tools built using the platform. Cluuz grew out of business intelligence software that Sprylogics designed to help companies mine and analyze data on their own internal servers.

The software was especially popular with financial regulators, such as the Investment Dealers Association, which uses Sprylogics’ technology to map the relationships and past dealings between companies and individuals it is investigating. To create a commercial search engine, Sprylogics simply took its enterprise search and analysis software and tweaked it to filter Yahoo’s search index data through the BOSS application. By delivering results that reflect what a user is actually looking for – rather than just locating keywords in certain pages – Cluuz and other semantic search properties believe they can also help deliver more relevant advertising to users than what Google currently offers. If, for example, someone were looking for information about installing blue headlights on their car, a traditional search for “blue headlights” would deliver pages that included those keywords.

A semantic search engine, however, would also bring up pages dealing with ‘illumination accessories’ or ‘after-market auto parts,’ said Tim Richardson, an e-commerce professor at Seneca College in Toronto and an expert in search technologies. It is that level of context that allows semantic search engines to present information in a way similar to how humans typically think. Semantic search is a hot area for both entrepreneurs and tech heavyweights.

In early July, Microsoft shelled out $100-million (U.S.) to purchase Powerset, a San Francisco-based company that produces a Wikipedia semantic search tool. Other semantic search providers such as Hakia.com – which has amassed more than $21-million in venture funding – are also growing their user bases while attracting interest from the big three.

Mr. Frank noted, ”We believe we’re one of the top three or four alternative search engines in the marketplace. You’re going to see more and more search engines become part of other applications, so we think that this application will be particularly well suited for when you’re looking for people or companies or entities of any nature.” With a market cap of about $5-million (Canadian), Mr. Frank knows that Google isn’t worried about Cluuz, but he does believe the standalone search engine his team has created can survive on its own or make a great addition to one of the top existing search sites.

Early semantic search trailblazers such as IAC/InterActiveCorp’s AskJeeves.com – renamed Ask.com in 2005 – have struggled to attract mainstream users since the mid-1990s, and have been met with a variety of other challenges, most notably cost. Developing software that can index the vast troves of information on the Internet, analyze it and then quickly present relevant search results is expensive and difficult. Last month, Cuil Inc. – a startup founded by ex-Google employees – unveiled a new search engine that claims to index three times as many web pages as Google.

But the site was met with an overwhelming criticism by users who found the results were often inaccurate and unhelpful. The key to success in semantic search is all about putting results in context. That becomes more difficult as the amount of information on the web increases, Mr. Richardson said, adding, ”The reason why people wanted to develop semantic web search engines is because it’s a more real way of finding stuff. Keeping in mind that the content is growing faster, search engines need to be more effective at reflecting the thinking of real people.”

Leslie Owens, an analyst with market research firm Forrester Research Group added, ”Some people think that the behaviour of how people express their information needs has been established by Google, so even though they don’t express it in natural language they do use just a couple of words. Can that behaviour be untrained if they could express things more completely in a full sentence or in a question? That’s to be determined. Google just has so much market share right now.”

Cluuz also shows the connections between various Web documents and sites based not on actual links between pages but on the information on the pages through a technology it calls semantic cluster graphs. It displays these results in a visual format that resembles a spider’s web.

Sprylogics International Inc. develops advanced search, analysis, and information display tools and services. These solutions enable users to search large amounts of unstructured data on the web, and in internal corporate databases, and convert it into actionable intelligence. The core technology driving Sprylogics’ solutions is embedded in the Cluuz Search Engine platform, which enables both consumers and corporate users to methodically search the Internet and internal corporate resources and find the information they are looking for. Cluuz search results are visually displayed through patent pending semantic graphs and result in improved decision making capabilities.

Source: Web Site Host Directory

CloudTuner – Poofy AND Stealthy

September 10th, 2008 by Rafi Farber
Posted in Newcomers, Reviews | No Comments »

Did you know that according to Yahoo!, 70% of us are typing in no more than 3 words in a search box? It seems that this threshold is something that’s wired in our brains rather than a limitation imposed by technology. Or maybe we just don’t have faith that adding more words will help. Given these three-word queries, it’s hard for a search engine to return what we’re really looking for. The situation is even worse when people are searching on mobile phones. Recent data indicates that 50% of queries performed on a mobile phone have a single term since typing more than that can hurt your thumbs.

Check out CloudTuner. (Demo videos showing how it works are at its website, though it itself is not public yet.) Its user interface comes to the rescue by providing an intuitive user interface for building more informative search queries linking many words together. Using the CloudTuner “term cloud” with embedded menus, a user can create longer queries while indicating the degree of relevancy or irrelevancy of each term. Within the CloudTuner’s menus there are built-in semantic services including spelling suggestions, related search terms, or language translations.

Time spent searching usually consists of analyzing results and refining a query accordingly. This process resembles the first HTML editors where content displayed during editing (HTML code) is different from the final output (Web Page Layout) so developers were forced to re-render their HTML code after each change to see how the final page layout is affected. These ancient tools are now replaced by new generation of software employing a “What You See is What You Get” (WYSIWYG) user interface for visual editing. Unfortunately, the standard user interface for search is still divided into query interface (search box) and results interface (list of links with summary).

CloudTuner takes a modern approach by providing a single paradigm for query refinement and analysis of the results. Users are allowed to shape the ‘term cloud’ into the desired form without going back and forth between a search box and displayed results.

Now we get to the really cool stuff: the image search. Image search is in second place in popularity on Google, but Google doesn’t look inside the images. It actually uses text found on the page around the image. There are dedicated image search engines actually analyzing image content and Google is rumored to have similar technology in its labs.

The way a content-based image search works is by looking for similar objects (in terms of color or shape) as in the image submitted by a user as a query. And here lies the problem: How does an image search know what object is of interest for a user? e.g. the car or the building behind the car? CloudTuner’s answer is to extend its cloud to image objects. It allows a user to refine results for zooming into an exact image piece (think man inside the car).

CloudTuner is Israeli startup currently operated in stealth mode. It’s working on releasing a Wikipedia search powered by its own technology. They plan to license their technology in the form of a software development kit to search engine vendors and directly to e-commerce or mobile portals. Stray tuned here on ASE for the latest demos and updates.