Not all search engines are created equal… but they mostly are.

“New search appliances claim to be uniquely adapted to meet enterprise needs,” writes Ben Dupont in Analysis: Enterprise Search for Network Computing magazine. “We tested eight enterprise search products and analyzed the technology's security and architectural implications. Our take: The math just doesn't add up.”

As I’ve written many times before, an ineffective search engine has less to do with the technology, and more to do with people and process and the ways and means by which content is categorized and tagged (see Fixing the sucky search problem and The search isn't broken, we're broken - Part I : Search and The search isn't broken, we're broken - Part II : Intranet Search & Taxonomy). This won’t however discourage the search engine companies from selling you.

I love the "live search" capability of the Plone search engine that we use on www.PrescientDigital.com. It starts to display results before you hit the "search" button.

Search results are automatically refined or updated in a drop-down menu as you type words into the search query box.

Insurance group Royal & SunAlliance has replaced its intranet search engine with Google's Search Appliance, which according to Silicon.com (see Royal & Sun Alliance Googles the intranet) has dramatically reduced search times on the corporate intranet.

“According to Royal & SunAlliance knowledge manager Tony Brierley, the replacement has reduced search times from between five and seven minutes to between 0.2 and 0.4 seconds.

The intranet is used by the company's underwriters to pull up information when dealing with brokers' enquiries. Underwriters require information such as rates or company policy guidelines that is stored on the intranet.

 

Brierley told silicon.com the company had a legacy of different search tools within the intranet, which is based on a Lotus Notes backbone. Interoperability problems between these systems slowed searches to an unacceptable level.” 

I’ve never heard of an intranet search engine taking 5 – 7 minutes to process a search query, but this is good salesmanship.

Ironically, Dupont in his analysis of the search engines is very complimentary of IBM’s OmniFind search engine, which is completely free, and obviously works well in a Lotus and Domino environment.

One engine Dupont didn’t review is the highly impressive Autonomy, which now eschews the search engine term as a simplistic label for its powerful IDOL server. Autonomy IDOL features “hyperlinking, agents, summarization, taxonomy generation, retrieval, channels, clustering, eduction, profiling, collaboration and alerting.” It denotes the difference between it and traditional search engines as “meaning-based computing.”

Autonomy distances itself from lesser search products by underscoring the key difference:  “keyword search engines for example cannot comprehend the meaning of information; these products were developed simply to find documents in which a word occurs.”

“Some of the key functionality of Meaning-Based Computing such as automatic hyperlinking and clustering are simply not available in keyword search engines,” states the Autonomy home page. “For example, automatic hyperlinking which connects users to a range of pertinent documents, services or products that are contextually linked to the original text requires that the meaning of the original document is fully understood. Similarly for computers to automatically collect, analyse and organize information computers have to be able to extract meaning. Only Meaning-Based Computing Systems can do this.”

All search engines were not created equal. However, notwithstanding some great marketing, salesmanship and the “Meaning-Based Computing” approach of Autonomy, they are remarkably similar.

RELATED READING:

Intranet vs Internet Search 

The search isn't broken, we're broken - Part I : Search

The search isn't broken, we're broken - Part II : Intranet Search & Taxonomy

Well Beyond the Search Box

Next generation inference engines

Analysis: Enterprise Search

  Digg this         Post to del.icio.us       Post to Slashdot   
 

  Add to Technorati Faves

Technorati Profile