Try Searching for Other Search Engines

Posted on July 2, 2006

Sure, Google is the most powerful and precise search engine around. It's now so vital to most journalists it has become a verb in most newsrooms, as in: "Google it to find out."

But there are other search tools around that, while not as comprehensive as the Google monolith, offer tweaks and tricks that make them worth checking out from time to time.

A9.COM OFFERS VARIETY

A9.COM is a multi-tasking search engine run by the same people who bring you Amazon. The standard search gives you pretty much the same results as Google.

But below its search bar, A9 also has various boxes you can check for additional searches:

Instead of just searching the web, you can click for a search of "Images", "Books" and "Reference" for example. So a search for "Winston Churchill" will get you the usual web pages - but also a list of books by and about the historical figure; pictures of him; and encyclopedia entries.

The advantage of this system is that sometimes you don't need the thousands of results Google will throw at you. You might just need the year Winston Churchill was first elected; or highlights of some of his famous quotes. Instead of drilling through pages and pages of Google results, A9 gives you all those results -- from the web, books and reference sources - all on the same page

CLUSTY GROUPS YOUR RESULTS

Many people confuse Google's speed with intelligence. Google is fast and precise, but it is not yet a thinking robot. Put in a search for "poverty" and "Canada" and Google will not return any pages about the poor in Newfoundland unless the word Canada is mentioned in that story. Because Google does not know Newfoundland is in Canada.
We are still a far way from the "Star Trek" computer that can collate and connect various seemingly unrelated topics.

But a company called Vivisimo is pioneering what is called "clustering" - grouping search results into folders of various sub-topics "based on textual and linguistic similarity." Their search engine at Clusty.com suffers from an ugly name and an even uglier logo, but it is a delight for certain kinds of searches. As they correctly boast on their web site, the clustering engine lets you see deeper and farther--with less effort--into a large number of search results to get a quick overview of the main themes that relate to the query.

Similar results grouped together for faster access and you can find results that are buried in the ranked list and would otherwise be missed.

Let's say you do a Google search on Ritalin and children. Google will give you thousands of results, ranked by relevance.

But put in the same two words in Clusty and you get a similar long list - plus a regrouping of the results on the left hand side. Clusty has thoughtfully assembled for you pages about parents of children using Ritalin; and schools and Ritalin; and about attention deficit disorder.

Clusty also has a "News" search function which performs the same kind of grouping.

So, for instance, a simple search for "nuclear weapons" gets you not just the latest news, but also a list of sub-categories, such as China, North Korea and Iraq.

The clustering results can be uneven. They work especially well for a topic about which you know very little and want to explore possible connections or trails of a story.

And sooner or later, Google and the other search engines are going to have to offer a similar kind of intelligence to help surfers wade through the mountains of results. One decent search engine that also clusters its results is called Gigablast.com

SPECIALIZED SEARCH TOOLS

There are also various specialized search engines that do not group or display results in different ways but look for information in different corners of the web.
Infomine at infomine.com looks for scholarly resources from journals and web pages, generally giving you a higher caliber result than general searches will.
HighWire Press at http://highwire.stanford.edu/ from Stanford University Libraries boasts that it is the largest repository of free, full-text, peer-reviewed content, with 926 journals and 1,323,421 free, full-text articles online.
Scirus.com targets only web sites with scientific content, drawing on resources from thousands of journals and books.

These tools also help peel away what is called "the invisible web" - the many web pages not usually found by traditional search engines. But that is topic for a future column.