Search Engines, Part 2 of 2

In the last issue of THE MEDIUM, this column generally discussed search engines. In this issue, search strategies and some lesser known albeit excellent search engines will be discussed. Uniform Resource Locators (URLs) for all the mentioned sites are alphabetically listed at the end.

Librarians may be especially frustrated when searching the Internet because they are used to knowing how to find information efficiently and effectively. A well-chosen keyword or two used to search an online catalog will find the materials in a library's collection. Yet those same keywords may not be as efficient when using a search engine. Library of Congress subject headings just do not seem to work as well on the World Wide Web. The reason is simple. The World Wide Web is a cooperative effort and international in scope. Anyone can create a web page and each page designer chooses the meta tags, i.e. subject headings, to describe the page. Much to the frustration of professional catalogers, this means that lay catalogers are creating their own hierarchies. Much of what information retrieval classes teach novice librarians does not apply to the World Wide Web.

While many library catalogs are moving toward natural language queries, most search engines have allowed natural language searching for some time. In the rarified air of the National Library of Medicine, a searcher using Medline products needs to know that the MeSH term for heart attack is myocardial infarction. At drkoop.com, however, the layperson's "heart attack" results in useful information.

Ask.com (originally called Ask Jeeves,) allows searchers to literally ask questions such as "Where can I find tourist information for England?" Ask.com provides web sites that may have some answers. The fact that ALL of the answers are not useful to ALL searchers is not a concern. One person may want information about walking tours of the Lake District; another may be looking for a West End theater guide. Both might find the information they want in the web sites offered as answers to "Where can I find tourist information for England?" and so the company considers the results successful.

It would certainly make searching much easier if all search engines operated alike. Sadly, that is not the case since search engines are designed and owned by competing companies. Some search engines require a plus (+) or minus (-) sign to require or eliminate terms; others allow for Boolean searching using AND, OR, or NOT. Some suggest adding quotation marks to indicate phrases. Some even allow a searcher to use a complex query such as "Indian jewelry"-Hopi+Zuni, which translates to "Find me web sites about Indian jewelry that include the word Zuni, but eliminate any that have the word Hopi in them." This is a dangerous proposition since some sites may discuss both Zuni and Hopi jewelry and by eliminating anything with the word Hopi, the entire site would be dropped from the results. Some search engines ignore any sort of indicators and. using the previous example, would look for each word individually no matter what the searcher does to indicate phrases. Such a search engine would probably offer web sites that have the word Indian (Indian fry bread), AND/OR jewelry (diamonds are forever), AND/OR Hopi (hairstyles) AND/OR Zuni (Zuni fetishes). The best advice is the simplest. When using a search engine for the first time, or when using a search engine that the searcher has not used in a while, spend the extra two minutes to read the searching tips. It might be called something as simple as "Help" or "Search Tips." Remember: Yahoo! will always be the best and most current source for information on how to efficiently use Yahoo!; Northern Light, the best source for information on Northern Light.

When searching for information about an historical figure or other famous person, HotBot (now part of the Lycos Network) is very useful. It offers a variety of menus along the left margin that allows searchers to limit the search in a number of ways. For example, one box offers the choice of looking for "all the words," "any of the words," "exact phrase," "the page title," "the person," "links to this URL," or "Boolean phrase." Further down, the searcher may choose to look only for images. Type in Winslow Homer, choose "the person" and check "images," and HotBot will find web sites with images of Homer's work.

Google is a highly rated searcher. While Google's refining techniques are not as obvious as those of some other search engines, it offers a button titled "I'm Feeling Lucky" along with the regular "Google Search" button. "Google Search" will return a list of web sites. "I'm Feeling Lucky" moves directly into the site at the top of Google's list of results. Type Andy Warhol into the search bar, click "I'm Feeling Lucky," and the searcher will be looking at the welcome page of The Andy Warhol Museum. "I'm Feeling Lucky" eliminates the middle step of following the link from the list of results to first site listed. If it turns out, sadly, that the searcher is not all that lucky after all and has been misdirected, it is easy to return to the original screen using the Back button on the browser. Google has recently added subject directories. Be sure to test your search skills with the "Google Quiz."

Northern Light is useful for scholarly research. It has a paid database, but also a free one and it breaks results down into folders, such as "commercial sites," and "government agencies" which expedites searching.

Looksmart offers a unique service. If, with normal searching, the query remains unanswered, the searcher may contact Looksmart Live (linked from the main page beneath the categories headings or via it's URL) and get help from a REAL PERSON. This is a terrific service for those having difficulty locating obscure information. I recently utilized Looksmart Live and received helpful assistance within 24 hours.

Possibly the newest old search engine is Lycos. About a year ago, Lycos changed from a simple spider to a portal, subject directory, and more. Overnight I went from never using Lycos to relying on it daily. Additionally, Lycos 50 (find the link at the bottom of the Lycos home page) offers a glimpse of searching trends. Lycos 50 daily examines different aspects of what searchers are looking for based on keyword searches performed on Lycos. Lycos 50 might be a useful tool to any searcher trying to think more like the Web. A searcher can try terms like "employment," "interview," and "resume," but Lycos 50 indicates that "Jobs" will work just as well.

By now most people are familiar with Dogpile. It is just one of many meta search tools which allows simultaneous searching on a number of different search engines. Similar search tools are: Highway 61, The Big Hub, Mamma, and Chubba. In addition to searching GoTo, Yahoo!, Alta Vista, Excite, Go (formerly Infoseek), HotBot, and Lycos (which is pretty standard for such search tools), Chubba offers a dictionary near the search bar to aid searchers with spelling

HotSheet is a general reference tool that offers searching in addition to ready links to a wide variety of sites. It is a fabulous resource; users might want to bookmark it.

I have just discovered two relatively new tools. I have not used either one very much, I want to give you're the opportunity to try them. MetaEureka allows the searcher to decide how many sites from each search engine to view, and also allows the searcher to weed out XXX sites. Ixquick uses a star rating system: one star for each search engine that has placed the site in the top ten results AND (listen up, librarians!) it allows truncation.

Two of the most common mistakes made by searchers are spelling errors and giving up too easily. If the searcher believes the topic is one that should get good results but gets few or zero hits, the first step is to double-check the spelling in the query. Even good spellers can make typographical errors that will cripple a search. If the spelling is correct, the next step is to consider alternative search terms. Using Lycos, I searched the term "arts & crafts architecture." One of the web sites returned was Arts & Crafts Internet Resources. From that site, I was able to cull other terms including Craftsman, Art Nouveau, and Mission Style, which could lead to useful web sites.

The last, and perhaps, the biggest hint is this: no search engine covers the whole Internet (the Internet includes FTP, newsgroups, World Wide Web, e-mail, Telnet, etc.). Each search engine actually indexes a shockingly small portion of the Internet (great ones may index as little as 16% of what is actually available). Search engines are competitive so there will be limited cross-linking. Just as CBS does not regularly promote NBC programs, it is unlikely that Yahoo! will direct a searcher to Snap's travel page or About.com's hobby page. If the spelling is correct, a variety of search terms have been used, and the results are weak, it is time to move on to another search engine. The whole world of information is out there; the trick is finding it.

http://www.about.com
http://www.altavista.com or http://www.av.com
http://www.ask.com
http://www.thebighub.com
http://www.chubba.com
http://www.dogpile.com
http://www.drkoop.com
http://www.excite.com
http://www.go.com
http://www.google.com
http://www.goto.com
http://www.highway61.com
http://www.hotbot.com
http://www.hotsheet.com
http://www.ixquick.com/
http://www.looksmart.com
http://live.looksmart.com (Looksmart Live)
http://www.lycos.com
http://www.mamma.com
http://www.metaeureka.com/
http://www.ncbi.nlm.nih.gov
http://www.yahoo.com