COPERNIC: Moving The Center Of Search Universe
The mastodons of online offer their users bland graphical interfaces, a handful of features, and super-size price tags. To their credit, the bluebloods of commercial online have embraced the Web browser as the sine quibus non of access for their customers.
But in Sainte-Foy, Québec, Canada (a tiny and often very cold Sainte-Foy), Copernic's 75 people have made state-of-the-art search-and-retrieval, real-time indexing, agent technology, and snazzy point-and-click interfaces something of an obsession. Copernic launched its newest version of its flagship software early in October 2002. Copernic is on a trajectory to add commercial content to their software in the near future.
Copernic has an approach to search and retrieval that has attracted a global customer base. Copernic reports that it has more than 20 million users worldwide and 250,000 users of the Copernic Plus and Pro products. That user base is roughly the equivalent of the number of people living in Canada.
Load the Copernic Agent 6 software, select a content area to search (there are more than 80 grouped under seven broad categories including business and economy), enter a query as a word, phrase or a natural language sentence, and start scanning results 10 to 15 seconds later. Updating the search engine data takes place automatically in the background.
The results are ranked by relevance, search terms highlighted, and the document summarized in well-formed sentences. Click a hit and the source document opens in the user's browser. Time delay? None.
Thoughtful, useful functions abound. The Copernic tool can be launched from within Microsoft Office applications or from the user's browser. A user can save the results of a search, schedule a Copernic agent to rerun the search at a time chosen by the user, or send the search results via e-mail in HTML format.
Users who want to limit a search to a topic domain click to select a category such as news, government and law, or business. Each of the seven broad categories can be narrowed. A click on the news folder shows such choices as business news, news in French, and nine other languages. If one of the predefined topics or clusters of sources is not suitable, Copernic Agent 6 allows the user to set up custom categories and save favorite searches.
Technically, Copernic is a type of search and retrieval software performing what is generally known as a "metasearch." Copernic and competitor Intelliseek allow a person to search multiple sources simultaneously. A documentalist trained in the use of Dialog's or Lexis Nexis's command line interface will know metasearch as a "411" or directory search. The idea is to streamline the search process by running one query or search statement across multiple sources. Copernic Agent 6 is to SS 411 as Mark Twain's typewriter is to Nokia's MMS mobile phone.
Intelliseek offers a competing product, Bull's Eye Professional. However, after receiving an infusion of venture capital from Nokia Ventures, Cingular (a U.S. wireless provider) and several other blue-chip firms shifted its focus to include a desktop metasearch product, an enterprise product, and a mobile device product.
There is another approach to metasearching. These are Web-based metasearch services that provide some of Copernic's and BullsEye Professional features from a public Web page.
One metasearch engine that has garnered a loyal following is Ixquick. Ixquick allows the user to restrict the query to the Web, music, images, and news. Ixquick's algorithms are based on statistical routines normally associated with ranking index funds. In addition, Ixquick uses a variation of the popularity scores pioneered by the now-defunct Direct Hit and search Titan Google. Ixquick is owned by Surfboard Holding BV, a Dutch company.
Other metasearch engines include the elegantly named Dogpile, owned by the now-asthmatic InfoSpace. A list of more than 100 metasearch engines is offered by Search Engine Hunter, although updates have been irregular. Another useful list is offered by Fujitsu.
The number of metasearch engines has been increasing as the number of Web search engines has been decreasing. The reason is that Google, FAST Search & Retrieval, and newcomer Wise Nut have driven less efficient, more costly spiders out of business.
As the search market has changed, Copernic achieved cult software status among many sophisticated research professionals. Anthony Safina, a legal and medical research specialist in Louisville, Kentucky, said, "I have used Copernic for several years. It allows me to make the most of my time, particularly when I am tackling an unfamiliar topic and when I don't know a specific source for a piece of information."
Not surprisingly, Copernic, after several years of development, entered the enterprise search market in February 2002. The firm's enterprise search competes with Intelliseek and a number of established brands in the enterprise search and retrieval market.
Why does Copernic think it can make headway in a crowded market? Search demigod Verity has teamed with such software vendors as Adobe to bundle search and retrieval into other types of software. Microsoft is bundling a search and retrieval tool with its new line of content servers. The wounded America Online provides the serviceable Personal Library Software as free, open source download for the budget conscious.
"A person researching technology from BEA Systems," says Mr. Bouchard, 20-something founder of Copernic, "can easily assemble clusters of Web sites from business news, newsgroups, technology reviews, internal content, and FAQs. When the query is entered into Copernic Agent, the results will be from these specific sites that our editors have included in the category."
But in many organizations, it often takes a miracle to find information. (It may be of interest to readers that frosty Sainte-Foy is named for the miracle worker in Liber Miraculorum Sanctae Fides.) Many portal and Intranet services offer search tools that are only partially implemented due to the large amount of editorial work that must be done on automated indexes and machine-generated classification systems. "It's a miracle if I can find what I need" echoes in the corridors at portal and search engine meetings.
Copernic's search results are deduplicated at different, user-selected levels of stringency. The user sees a list of hits with a 30 to 100 word abstract. Each search term can be highlighted in one color or different colors for each term.
Copernic is a privately-held company and does not provide specific data about its revenues. According to Mr. Bouchard, "the company is profitable."
Says Mr. Bouchard: "We have put agent technology in our enterprise search product. This allows any user without any training in search whatsoever to locate information anywhere on the corporate Intranet or available via the World Wide Web, see summaries of these documents, and save the search so it can be rerun anytime or automatically."
Copernic has learned that the weakness of most of the search and retrieval products aimed at the enterprise market come down to a handful of problems that users readily identify.
He says, "We have engineered our indexes to allow the customer to build a single index and update it in real time or use multiple indexes to meet special security requirements. Updating and refreshing indexes is the big problem for many enterprise search engines. Believe it or not, some well-known products have to reindex the entire domain of content just to refresh the index. This approach makes systems professionals quite frustrated because overall network performance is severely affected."
"A second problem," he says, "is delivering advanced features such as our interface, support for the many different types of files and documents in an organization, and precise Web access at a reasonable price. We charge US$12,000 per CPU. There is no charge for the number of users or the number of documents indexed. Our pricing model is easy to understand and budget."
Mr. Bouchard notes, "The number of search engines focusing on specific content niches is starting to rise. In 1999 there were maybe 75 to 100 of these services. Now there are thousands of high quality sources. A good example is Moreover.com, a service that focuses on news from well-known sources and Web sites that update frequently with original content." Other examples of vertical search engines include the infamous KaZaA and such specialized services as Search Security, a site providing an index of computer security content.
Copernic has continued to expand its product line. In addition to the enterprise version of its software, the company offers a tool to track product prices, an effective, fast, stand alone summarizing tool that fills a void left by the Australian development company InText, and a new version of the company's flagship product.
Copernic Agent Version 6.0 was released early in October 2002. It includes additional automation functions. Mr. Bouchard says, "I dislike phrases such as 'neural networks' and 'artificial intelligence'. We prefer the term 'agent' or 'robot.' In Copernic Agent 6 and our enterprise product, the implementation of 'agent' is pragmatic. An 'agent' is a software that does work for the user. At Copernic we try to identify work tasks that can easily be given to an agent who will perform certain tasks that users find repetitive or tedious."
Copernic Agent 6 and the Empower enterprise product present a well-organized interface. The young at heart will revel in the visual sizzle of the five panel, multi-pane screen layout--search progress indicators, search options, highlighted results. Those with a foot in the design camp of IBM green screens will be disappointed.
Most users will be clicking and searching moments after loading the software. Usability gurus will grouse about the amount of information presented. However, Copernic--unlike Adobe's software content management products-- does not hide important functions or change the state of the software with each mouse click. (Adobe, are you listening?)
Mr. Bouchard says, "Copernic's products have a common code base. Our products are completely object oriented and optimized for speed and scalability. The development tools we use give us considerable flexibility when it comes to customizing the interface for a particular application."
Copernic has another wave of innovations scheduled for later this year or early 2003. Among the projects that are nearing commercialization are enhanced relevancies. "We continue to work on our relevancy algorithms. We have a team experimenting with the newest artificial intelligence, neural network techniques from university researcher departments. We search for more effective ways to use linguistic cues and outputs from statistical routines.
Copernic's research and business development professionals have accelerated their work to provide a common interface for Web and for-content. Mr. Bouchard notes, "People looking for information do not want to use one interface for the Intranet content, one for the Web, and then more interfaces for content from established publishers. We are finalizing the Copernic tool to handle all these content domains from our software. In a near future, we see the user in the Copernic interface and receiving a single invoice for the for-fee content retrieved and downloaded."
Compared to search and retrieval systems firmly rooted in the 1960s, Copernic's approach makes clear the distance between the mastodons of online and the miracle workers in Sainte-Foy.
Web site mentioned in this column include:
Copernic: www.copernic.com
Intelliseek, www.intelliseek.com
Stephen E. Arnold
President
Arnold Information Technologies
Postal Box 320
Harrod's Creek, Kentucky 40027
Voice: 502 228 1966
E mail: sa@arnoldit.com
Web site: www.arnoldit.com
Mr. Arnold is an independent consultant working from Harrod's Creek, Kentucky. He resides in the USA with two boxer dogs and two new Xeon servers and his trusty sidekick Benny, a four processor Netfinity. His newest book is Knowledge Management Sense and Nonsense: How to Make Data Payoff. It will be published by Infonortics, Ltd., Tetbury, Glou. in early 2003. Mr. Arnold specializes in business strategy, competitive intelligence and technology analysis.
The Copernic Agent 6 display shows from left to right the primary search box, previous saved searches, results, page preview, and common tasks. Buttons and icons feature balloon help. The display can be reconfigured to suit the user's personal preferences.
[ Top ] [ AIT Home ] [ Beargrass ] [ Site Map ]