Exalead presents Wikipedia search, blog and forum search and an improved image finder.

By Pandia Guest Writer Lars Våge

The French search engine Exalead has probably the most comprehensive set of search functions of all search engines available. Regular searchers do not concern themselves with advanced search functions, which is probably why they haven’t heard about Exalead. That’s a shame!

Exalead has presented several new features lately:

Restrict to blogs

You may now restrict the list of search result to blogs or forums. In the search engine result page you will find a box called “Refine your search” on the right hand side. This new feature is found under the subcategory “Site Search”, marked as “Beta”.

You may also restrict your search to file type (e.g. PDF), media files (images, sound, video), language, geographical location and related searches. You may select one or more of these options to narrow down your search.

Unfortunately, the algorithm used to identify blogs and forums is less than 100 percent accurate. It is closer to 70 percent, so there is definitely room for improvement.

New interface for Wikipedia searches

On Friday the 13th Exalead presented a beta version of a Wikipedia search interface. I hope they will succeed in spite of the launch date.

I like Exaleads user interface very much, including the options for refining searches. However, the Wikipedia search “Narrow your search” box is a little bit different compared to other Exalead searches: This one has a tag cloud.

I find the tag cloud pretty useful (which is more than I will say of other tag clouds). At the top there are flags symbolizing restrictions to English, German, French, Italian, Spanish and Dutch articles. Below the flags you find regular tags, where the words with the largest font sizes represent the most popular hits.

Exalead’s tags are also colored, which is unusual: Redbrown represents categories, dark blue: related terms, orange: persons, bright blue; geographical names and grey: organizations.

Under each Wikipedia article search results you will see the categories associated with that article and links to articles presenting persons mentioned in the article. There are also links to organizations and geographical names linked to in the article — however, this requires that you activate full view by clicking on the relevant icon to the far right besides “View”.

You may also use the “View” option to get a more Google like interface (i.e. without the preview thumbnails) — but why would anyone do that?

Exalead does have some difficulty in identifying geographical names. If you — for instance — search for “Eric Dolphy “and click on results you will find a few errors. “Ascension” is a record by John Coltrane, “Point of Departure” an album by Andrew Hill and “Gongs East” by Chico Hamilton. These are not geographical locations.

In Exalead’s Wikipedia search there are several advanced search functions found in the regular Exalead Web search engine, but which are missing in Google, Yahoo and Microsoft’s search engines. These include the Boolean operator NEAR, truncation of words using an asterisk (*) and the uses of nested parenthesis in Boolean searches. The NEAR proximity search is the most useful of these features.

Image search

Exalead has also improved the image search feature, apparently based on feedback from users. This is all about searching for images and then restricting results to images containing faces. Exalead has been helped by technology delivered by LTU Technologies, a company that specializes in face recognition.

If you want to test this feature you do an image search from the Exalead home page. On the first page of results click on “Narrow your search”. Then restrict the search by selecting “Face” under “Content”. How many of these pictures will really depict faces? I have done some testing and find that some two to five out of twenty search results do not contain faces. Not too bad, I can definitely live with that.

The Exablog

Note that Exalead has launched its own developer blog in the same tradition as the larger search engines.