15 June 2005. As much as I talk about how to use information discovery tools beyond search engines, once in a while I find a new search tool that makes me actually want to do some Web searching. Exalead is a fairly new search engine from France, which was introduced in October 2004.

It is still officially in beta. Having passed the one billion page mark in 2005, it's still one-eighth the size of Google or Yahoo. But what's a few billion pages among friends? Actually, after a certain point, size really doesn't matter. The key factors in evaluating a search engine should include timeliness, ability to handle ambiguity, and plenty of power search tools. Exalead does a great job, at least on two of these criteria.

Exalead enables true truncation, where you can search for "librar" and retrieve library, libraries, librarian or librarianship.

When you first connect, you see a stylishly minimalist page. But click through to the Advanced Search page to appreciate Exalead's search features.

Among the features, which you don't always find in search engines, are:

• The option to specify that results "preferably contain" all the terms you are searching for, in addition to "must contain" and "must not contain."

• You can also do this with the OPT operator, to indicate which specific words are "optional."

• Proximity searching, in which the words you search must be within 16 words of each other. (No, you can't tweak the number of intervening words.)

• Truncation, and this isn't just the word stemming that many search engines employ behind the scenes (a search for "pencil" will also retrieve "pencils"), but true truncation, where you can search for "librar" and retrieve library, libraries, librarian, librarianship, and so on.

• Phonetic spelling and approximate spelling, through which you can search for a word, even if you aren't sure of the spelling, or if the word is frequently misspelled. Think "Arnold Schwarzenegger" for example.

• What Exalead calls "Regular Expressions," in which you can search for documents with words that match a certain pattern. Imagine, for example, that you're doing a crossword puzzle and have a word of 6 letters, of which the second is T and the sixth is C. By searching /.t...c/, you will retrieve sites with the word ATOMIC, perhaps the right word for your puzzle.

One gripe I have is that there is an option to limit your search by country but, unfortunately, this only searches by two-letter top level domain (e.g., .uk, .jp). This means that, for example, if you limit your search to Australian sites and search for Australian biotech associations, you won't retrieve AusBioTech.org, a major biotech association in Australia, because it does not have .au as its top level domain.

In addition to search power, Exalead has a rich search results screen. In addition to the usual display of search pages and snippets, each entry includes an image of the retrieved page. There is also a column along the left that displays relevant entries from the Open Directory Project, along with tools to select "related terms," to limit your search by document type, and to narrow the search by location (and, interestingly, this doesn't use the two-letter top level domain limit, but instead retrieves only pages from the Open Directory Project that have been categorized under that country.)

My one real objection to Exalead -- and it's a big issue –- is that it appears Exalead has not updated its index since the beginning of 2005. One of its advanced search features lets you limit your search by the date a file was last modified (note that you need to use the European format of dd/mm/yyyy). But repeated tests turned up no records from 2005. Yes, Exalead is in beta, and that sometimes means there are glitches, but a delay in updating the index is troubling.

Until Exalead gets its updating schedule back on track, use the search engine to find older material, or to verify spelling, identify alternative word meanings or find authoritative material from sites that have a track record. And show this site to the next representative of one of the value-added online services. The features in Exalead would add tremendous search power to, say, Dialog, Factiva or LexisNexis.

© 2005 Mary Ellen Bates all rights reserved.

Mary Ellen Bates is the principal of Bates Information Services, a research and consulting business based in Boulder, CO.