The topic of this week's feature article is Web-based search project that tries to solve the ancient problem: how to search
for Web resources, identify valuable information, and access that information efficiently, when the number of Web sites is doubling
every 6 months. OK, there are a dozens of search engines,
and there are many full-text indexing systems for intranets, but they all take a very mechanistic, nonintuitive approach.
Meet Alexa, created and designed by Brewster Kahle, inventor of the famous WAIS (Wide Area Information Server) -
an important precursor to Web directory and indexing sites. After selling it to America Online for 15 million dollars, Mr. Kahle designed
new "search engine" that doesn't rely on word searches, but makes it much easier to search electronic data bases.
Alexa is software that can be retrieved free from the company's Web site
and added to a Web browser. You don't have to type keywords or search phrases: instead, Alexa watches where its users go
on the Internet, and then records that information in a central data base. Based on that information, it can tell a user the most
popular paths that other Alexa users have taken from the site the user is visiting at a given time (Where to go next?).
It also can suggest other sites
offering related material. The top 10 sites pop up in a toolbar near the browser and change as the user moves from page to page.
Picture taken from the Alexa Internet Web site
Alexa offers instant background information on the site you are currently visiting so you can determine the value of that site
(Where am I?). Whenever your browser goes to a Web page, the Alexa toolbar requests information about that page
from Alexa's servers. Server then records that an Alexa user has spent time at that site as a kind of vote - it doesn't matter who
passed through this site, but server "knows" how many have passed this way. It can also learn from the fact that users visit
several related sites as they browse the Web, and tend to spend more time at the sites that give them more of what they are
Users can also explicitly make suggestions of Where to Go Next from a given Web page.
They do this by selecting "Add a link to this list" in the "Where to Go Next" pop-up on the Alexa Toolbar. These signs
are always visible to the user who placed them---they are like context-sensitive bookmarks (Netscape Navigator) or favorites
(Microsoft Internet Explorer). But they are also used when Alexa tries to determine which sites to suggest to the new users.
When using Alexa toolbar, it is also easier to access the leading reference tools like Encyclopaedia Britannica Online and
Merriam-Webster's Dictionary and Thesaurus. Your desktop reference tool is always only one click away no matter where you
are on the Web.
Alexa's "Where to Go Next" suggestions are influenced by the paths users take and the suggestions they make.
To have something useful to say before it has lots of users, an analysis of the text of the Web (finding pages that talk about
similar things) with an analysis of the structure of the Web (what pages point at what other pages) is performed.
Alexa's "Where You Are" pages are based on correlating a wide variety of public databases with the information Web site
publishers give when they register their site names.
When Alexa user gets "404 - File not found", they can click on a button on Alexa's tool bar and pull up
the missing page from the archive (What happened to my page?). As of December 1, 1997, Alexa's Web
archive is in excess of eight terabytes (a terabyte is a million megabytes!) and the new snapshot is taken approximately every
30 - 60 days. Alexa began collecting the Web in early 1996 and now has at least three-four snapshots of over
500,000 Web sites. It is beginning to adjust its visits of sites to reflect which sites change most frequently.
It is easy for Web site administrators to request that their site be archived at a particular time or not archived at all.
The company is located in The Presidio (Building 116) of San Francisco, California. You can reach them at
http://www.alexa.com, or, by e-mail, at