Fortunately, we have another device in our Internet bag of tricks. WAIS (pronounced "ways"), short for Wide-Area Information Servers, digs beneath the titles of Internet documents to search their content and return files that match our criteria, regardless of their names. WAIS lets you look through a large pool of information and helps you find something even if you aren't exactly sure where it is. If Gopher serves as the Internet's table of contents, WAIS acts as its index.

WAIS was begun at Thinking Machines Corp. in the late 1980s and has undergone many changes since then. Developer Brewster Kahle and others founded WAIS Inc., in 1993. The company sells commercial versions of the WAIS server—the product that businesses use to gather all their documents in text, audio, and video form to create a searchable database. But for our purposes, what's important is that the client end, or the browser that is used to conduct WAIS searches, remains free of charge and is likely to stay that way for some time.

Let's assume we want to search WAIS databases for information on Asia. First we find a WAIS client, and then we type in our query using the keyword "Thailand." WAIS searches through the hundreds of databases and returns with a listing of documents that have to do with Thailand—travel advisories, Thai history, and bibliographic references to Asian religions.

Here's how WAIS works. You make a request to search through all the databases that are on WAIS; at last count, there were an estimated 500 or so WAIS databases, with titles such as zen.internet.src (Zen and the Art of the Internet) and sfsu-phones.src (San Francisco University telephone directory). WAIS then goes through all the documents on those databases and looks for matches to your request by asking each server to search its index.

Type in a keyword, and WAIS is at the ready to search through hundreds of databases to come up with a list of documents that relate to your keyword.

A few seconds later, the server returns a list of documents that are ranked according to the number of times the keyword appears in them. Once you've pinpointed the database(s) you want to search, getting a specific file takes just a few seconds. WAIS limits the search to the most relevant documents; depending on the client, you may see up to 50 hits.

Powerful Technique.

WAIS uses a search technique called relevance feedback, one that is more powerful than the traditional Boolean search. With a Boolean search, you use connectors such as OR, AND, and NOT to retrieve documents. For example, if you want to find articles on house pets, you might type in the phrase "cats or dogs," and the search would return articles that refer to cats or dogs in the text. You then could narrow your search by adding "not goldfish" to cut down on the number of documents you receive, but you can't take the articles you found most pertinent and use them to narrow your search.

WAIS, however, does exactly that. It takes the results of your search, feeds back the most relevant files or articles, and tells the server to find similar files. It's important to keep in mind, too, that if you were to give a WAIS client the same keyword string you might give a Boolean search engine, e.g. "cats and dogs," it would return documents that featured the word "and," so be sure to avoid mistakenly typing connectors.

Another reason WAIS is such a powerful client is that it searches through the text of documents and not just their titles, as Veronica does in Gopherspace. Let's take the example we used when describing the limitations of Veronica. Suppose you're interested in finding articles on Bosnia and you use that as your keyword. A Gopher menu titled "The Land Struggle of Croats, Serbs and Muslims " would not be turned up under a Veronica search, but because WAIS looks through the entire text of documents, that article would appear with a WAIS search. And another of WAIS' strong features? It's free and open to all users. Unlike Telnet, where users may need to enter a password to access a remote site, WAIS databases are there for all explorers.

But WAIS is not perfect. It does not turn up as much comprehensive information as searches of other sections of the Internet might, because WAIS searches a limited number of databases. You probably would find a lot more information about the U.S. government with Yahoo!, because the World Wide Web is such a large publishing platform.

Another drawback is that WAIS scores documents based not on the frequency, but on the number of times a keyword appears, so longer documents have a better chance of appearing at the top of your list of hits. And WAIS returns a database that references document titles without summaries, so until you open "com-books.src," you won't know what's inside.

The Means To WAIS.

So let's look at how you can go about using WAIS. If your Internet access consists of going online via a commercial service, you won't be able to use any fancy graphical tools to conduct WAIS searches. Neither CompuServe nor Prodigy has special WAIS sections, and although America Online has an area for Gopher and WAIS, the WAIS engine is not yet functional and technical support could not give us an estimated date of arrival.

A simple way to use WAIS is to use Telnet to find a site that runs swais. Depending on your interface, log into the Telnet site by using Telnet to find quake.think.com and log in as "wais" to get to the WAIS server at Thinking Machines Corp. (For details on Telnet, see the related article in this section.) From there, you can type in keywords and run the same searches you can run with a graphical client, but the interface is somewhat challenging and recalling the commands can be daunting. You also can access WAIS by using Telnet to find the sunsite. unc.edu site, which prompts you to enter "swais" for a simple WAIS client, "politics" for a WAIS search of political databases, and "linux" for a WAIS search of Linux databases (those that deal solely with Linux, a Unix-based operating system.) Once you are logged in, the commands at the bottom of the screen will tell you how to type in a keyword and conduct a search.

Another option for using WAIS, one that is more fun than telnetting to a WAIS server, is to use one of the many graphical WAIS clients. The first step is to find them, which is easy if you know how to use FTP. (For more information on using FTP, see the related article in this section.)

We used FTP to retrieve the WinWAIS client, an easy-to-use graphical interface. After retrieving the file ewais200.zip from ftp://ftp.winsite.com/pub/pc/win3/ winsock, we placed the file in a temporary directory and decompressed it using PKZip. (For more about downloading, or decompressing files, see related story in this section.) Then we set up the program, clicked on its icon, and watched as the file came to life. (Because this process differs from application to application, we won't go into the details of FTP here. However, for FTP novices, we recommend America Online's FTP capabilities because of their intuitiveness and ease-of-use.)

Running the search is now fairly straightforward. Just type in a keyword or keywords under Tell Me About, click Ask, and watch WinWAIS go to work. If you see a file you would like to read, click View or Save under Retrieve Document, and when you're ready to log out, you can ask WinWAIS to save this source to be used in future searches.

By now you should have a solid understanding of WAIS basics. Here are some valuable resources to help you keep up with the latest WAIS-related information:

•comp.infosystems.wais. This Usenet newsgroup comes close to being a technical support line for WAIS users. Users post articles asking how to get formats to work with freeWAIS (a freeware interface for using WAIS), how to create an index with user-defined fields, and how to set up MacWAIS (a graphic interface for WAIS on the Macintosh).

•quake.think.com. Use Telnet to access this site operated by Thinking Machines Corp., where WAIS was first developed, for software and documents.

by Heidi Anderson