appearing on the AIMR Web Site at <www.aimr.com>

Search Engines

The title "analyst' implies someone who is acquiring information, studying and balancing relationships to reach judgements. It is what we in the AIMR do.

We can all remember the wonder of the first day when, as children, we learned that libraries held so much knowledge. We explored the stacks for hours, days, or more to find what we wanted. We learned the secrets of the Dewey Decimal System to speed up the time we'd spend with those heavy (for a kid) wooden file drawers. And we were awe-struck when the reference librarian could instantly direct us to the sources we wanted but could not find on our own.

There is no really accurate estimate of the amount of material on the web today. It might be one hundred million pages but with changes, links and growth it could be more or perhaps less. It is an incomprehensible library, a resource of our dreams. Its only problem is that it is too much; it needs to be more personal so we can navigate quickly to just what we want.

Search engines became the tools of choice. Perhaps Yahoo was the first, most popular engine. Like many of the best early things that happened to popularize the web, it was essentially a pet project of a couple of California student wits who called it "Yet Another Hierarchical Officious Oracle." By going to Yahoo , you could put in a few key words and find most of the documents on the web with one or another of the words. You could get an exact match by enclosing your words in quotation marks and further narrowing your search with so-called Boolean rules (most search engines follow this convention and give some guidance on their use). Yahoo got so big and unwieldy that it began to segment itself into categories, rather like the sections of a library. And it became the most often hit of the web sites while supporting itself, no long a part-time project, by advertising. Infoseek operates without Boolean rules and provides a slightly different list of search results. Analysts will tend to try one search engine after another if the first attempts are not successful. Each effort tries to get a different facet of the search question being pursued. Lycos is another common engine that also gives a slightly different list.

And Yahoo faced competition. Each new tool had some benefits. Magellan would grade sites on their potential usefulness, but that was rather limited. Excite uses some artificial intelligence to refine searches and provide a rich pattern of other related sites with a "more like this" suggestion when you find something of interest. It takes you on to more sites of the same branch.

AltaVista is reputed to give the broadest search. And recently it has added a beta of automatic web page translations that works very well through a link on AltaVista's home page. Automatic translation is especially useful for analysts studying countries that use their native language (which is non-native to most of us).

There are many other search facilities. Some are targeted within popular sites appealing to special interests. A number are useful to us, including Ohio State's Virtual Finance Library. Then there is a new category of meta-search engines that search a number of engines at the same time. I am a strong devotee of Copernic that searches twenty popular engines, including news groups, for matches and combines the results into clear and graded summaries in its latest Preview 98 version.

We see early indications of the future of search engines. Firefly matches an individual's characteristics and pattern of web studies to those of others and suggests ones that might be personally interesting. Similarly, Alexa collects "votes" of sites from similar people as a tool to guide users to sites which may have appeal. We shall see intelligent agents as proxies for each of us that go through sites looking for new material, providing ratings that match our individual preferences…all without our direct intervention.

Search engines are important because of the frequency that we look for information. Information sources are growing so rapidly that, to conveniently find and sort what we want, search engines-most likely as intelligent automated agents-are our third hand.

Dean LeBaron
March 5, 1998

email <deanlebaron@compuserve.com>

website <http://www.deanlebaron.com>