I love search engines.
Last time I wrote about them, four years ago, I recounted the story of how I looked up ''Easter'' and came up with everything but the holiday -- including recipes for Easter pirogi, a guide to Easter Island, and lots of references to easter eggs -- not the colored kind you hunt for, but the slang term for hidden features in computer software and hardware. Easter, the holiday, showed up as roughly the 100th entry in some of the top engines.
Well, progress marches on, as the technology folks continue to refine their algorithms. I looked up Easter on Google, (google.com) this year's hot search engine, and got ... easter eggs (the computer kind), Easter Island, Easter Seals and, finally, in fourth place, ''Easter on the Internet'' (holidays.net/easter/), a genuinely useful reference. Hotbot.com and Yahoo.com did better, with virtually all their references pointed to the holiday.
I run little tests like this every now and again. Search engines overall are getting better and some work better than others depending on the query. Still, there are some caveats:
1. The Internet does not contain the sum total of human knowledge, and search engines can't find what isn't there.
Information that's easy to find in just about any public library may not be on the Internet. A search engine, unless carefully directed, may be less useful than a good encyclopedia.
2. A search engine is more likely to give you too much information than too little.
Looking for a very general topic with a search engine is going to produce many more answers -- most of them wrong -- than you can possibly read. And there is no guarantee that the information itself is accurate.
3. You can look for information in one of two ways: through hierarchical lists compiled by people or through indexes compiled by search engines.
The former are more useful for broad topics, while the latter are used when you want a narrow, highly specific piece of information. Metasearch engines query more than one search engine, then sort, sift and rank the results.
Pay attention when you're on Yahoo next time and you'll see what I mean by hierarchies. The opening screen offers a search engine (which covers the entire Yahoo database), along with topics ranging from Arts and Humanities to Society and Culture. Click on ''People'' under Society and Culture, and you'll get more subtopics ranging from American history to women; click on ''Women,'' and you'll see more topics, subtopics and so on.
Search engines look at the Internet as a whole. The basic scheme of a search engine works like this: It examines a Web site and builds an index of key words relating to the site. These all get stored in a central database, which is what you are querying when you do a lookup. Obviously, search engines can't examine or index all sites at once; rather, they ''crawl'' the Web for hours or days compiling their indexes. Search engines may miss sites altogether or update the index of their contents in a tardy fashion. Search engines also differ in how they decide what's relevant to your query.
4. Different search engines produce different results.
Looking up even an uncommon name such as ''Dolinar'' can produce hundreds of hits, which take forever to examine, so engines are programed to rank responses in order of relevance according to rules that aren't consistent from one engine to another.
5. Lots of the good stuff has been sectioned off and tucked away in proprietary data stashes.
Newspapers and magazines, for example, may keep their clip libraries in places where they can't be accessed directly from a search engine -- you'll have to go to a specific site, set up an account there (which may or may not be free) and search the site. That's what you have to do with Dow Jones and the Wall St. Journal. Other papers -- Newsday, for example -- will let you look at recent data for free but charge you to access older archives. Another common form of ''payment'' simply has you register to obtain an account and password; this helps the site sell ads because its owners can say who is reading.
There's a lot of useful stuff available for free on the Internet.
Unfortunately, the more valuable the information, the more likely it is to 1) cost money and 2) be hidden from search engines.
6. Specialized databases are usually preferable to general-purpose search engines.
A good example is shareware databases, which we wrote about a few weeks back. If you're looking for information about computers, you're better off scanning the sites of computer magazines than querying the Internet at large. Take a look at crawl-it-all.com for a super page of specialized search engines.
Is one search engine best? I've seen tests and reviews that purport to answer this question, but they rarely agree. Take a look at the best single source on the Internet, Danny Goodman's Search Engine Watch (searchengine watch.com/), which has clipped just about every report and magazine article ever written about the effectiveness and popularity of the various sites. There's also a ton of original material on how to search, and how the various engines go about finding what you seek. A must-read, as they say.
I go through phases with search engines, and search engines seem to go through cycles, too. AltaVista used to be comprehensive and wonderfully fast, but as more users have piled in, I find my queries sometimes hang for minutes at a time, so I use it less and less. MetaCrawler's pretty good, too, and Hotbot is usually worth a look, but right now I mostly use Google, which is lightning fast.
Yahoo rounds out my collection.
That's me. You can get other insights by looking at which services are the most popular. The top services, roughly in order depending on how you're counting, are: Yahoo, the Microsoft Network, Lycos, Netscape, Infoseek, Excite, AltaVista, Snap, LookSmart, AskJeeves, GoTo, iWon, Go2Net, and DirectHit.
Distributed by the Los Angeles Times-Washington Post News Service
Brainerd Dispatch ©2013. All Rights Reserved.