WAIS FAQ

What is WAIS?

WAIS, Wide Area Information Server, are databases containing mostly text-based documents, (although WAIS documents may contain sound, pictures or video as well). WAIS databases are referred to as sources. The databases may be organized in different ways, using various database systems, but the user is not required to learn the query languages of the different databases. The WAIS client uses natural language queries to find relevant documents and the result of a query is a set of documents which contain the words of the query. WAIS databases are available on topics ranging from Agriculture to Social Science.

Who can use WAIS?

You must be on the international TCP/IP network (the Internet) in order to use a WAIS client on your computer to access WAIS sources.

If you have e-mail access only, or if you are not on the Internet, you can still exploit some of the potential of WAIS. An interface which gives e-mail access to WAIS databases is available at the address: WAISmail@quake.think.com (see E-mail access section under Using WAIS below).

How to get to WAIS

There are many WAIS servers throughout the network. A directory- of-servers database is available at several sites, which can be queried to find out what databases are available on a particular subject. The directory-of-servers database is also available via anonymous FTP from ftp.wais.com in the directory /pub/directory- of-servers in the distribution wais-sources.tar.Z as file directory-of-servers.scr.

If you do not have access to a WAIS client, at least two demonstration sites are available to allow you to get acquainted with WAIS. You can telnet to:

       +---------------------------------------------------------------+
       |                                                               |
       |  quake.think.com  (login: wais)                               |
       |  sunsite.unc.edu  (login: swais)                              |
       |                                                               |
       +---------------------------------------------------------------+

The two demonstration sites above run swais (Screen WAIS), a simple WAIS client for Unix.

Using WAIS

There are many freely available WAIS client programs for various operating systems (Unix, VMS, MVS, MS-DOS, OS/2 and Macintosh) and for specific environments (e.g. X-Windows, Openlook, NeXT, and MS-Windows). See the list of freely available client software in Appendix A.

The client interface differs slightly on different platforms, but the queries are performed in the same way whatever interface is in use.

Step 1: The user selects a set of databases to be searched.
Step 2: The user formulates a query consisting of keywords to be searched for.
Step 3: When the query is run, WAIS asks for information from each selected database.
Step 4: Headlines of documents satisfying the query are displayed. The selected documents contain the requested words and phrases. Selected documents are ranked according to the number of matches.
Step 5: To retrieve a document, the user simply selects it from the resulting list. The WAIS client retrieves the document and displays its contents on the screen.
Step 6: If not enough documents are found, the user can state the question differently or can select a set of documents.
Step 7: A further search will look for documents which have a large number of words in common with the selected documents.

E-mail access:

You can query WAIS databases and retrieve documents by sending commands in the body part of an e-mail message to WAIS-mail@quake.think.com. The Subject: line is ignored. The important commands are given below (a vertical bar (|) indicates a choice of parameters):

help: to get the help file
maxres number: to set the maximum number of results (i.e. documents) to be returned.
search source-name | "source-name1 source-name2 ..." keywords where: source-name is a source name as found in the directory-of-servers (with or without the .src ending). Use double-quotes (") to group several sources to be searched. keywords are the words you would normally type into a query. You may specify several search requests in a mail message. If you don't know what sources you can search, just try anything. If the source name is not recognised, you will be given a list of sources.
retrieve docid: retrieves a document from a database. docid is a Document IDentifier as returned by a search. You may put more than one retrieval request in a mail message, but you must leave a blank line between requests. The docid must be written exactly as returned by a search request, including any spaces. You can retrieve non-text documents as well as text. If the document is of type TEXT or WSRC you will get the result directly. Other types will be UUENCODED before being transmitted.
DocID:: docid same as retrieve. This form is identical to the form which is returned by a search request, which makes it easy to use the reply mail function to retrieve results.

Examples

When you log in to the demonstration site at quake.think.com, you have immediate access to the directory-of-servers database via the swais client software. To find recipes using papaya, you would select the recipes database and give papaya as the keyword. Here are the results of the search:

       -----------------------------------------------------------------

         #  Score Source                  Title                    Lines
       001: 1000 (recipes) arielle@ta Re: Dawn's Muffins, Pt III     339
       002: 1000 (recipes) arielle@ta Re: Muffins 3                  632
       003: 1000 (recipes) arielle@ta Re: Pineapple                  678
       004:  750 (recipes) arielle@ta Re: Pork and Papaya Salad       33
       005:  750 (recipes) arielle@ta Re: Bread                      681
       006:  500 (recipes) roder@cco. Re: NONFAT BAKERY COLLECTION   423
       007:  500 (recipes) shiva@hoss Re: Juice Recipes               65
       008:  250 (recipes) arielle@ta Re: Prawn Salad                 49
       009:  250 (recipes) arielle@ta Re: COLLECTION: Lots of Avoca  447
       010:  250 (recipes) mecca@acsu Re: REQUEST: blender-made fru   29
       011:  250 (recipes) Ann.Adamci Re: Re: REQUEST: blender-made   38
       012:  250 (recipes) patth@Pani Re: Re: REQUEST: blender-made   49
       013:  250 (recipes) arielle@ta Re: Avocados                   459
       014:  250 (recipes) red_trek@d Re: VEGAN: red beans and rice   78

       -----------------------------------------------------------------

You can then select any of the above documents for viewing. Here is the Pork and Papaya Salad recipe:

       -----------------------------------------------------------------

       Newsgroups: rec.food.recipes
       From: arielle@taronga.com (Stephanie da Silva)
       Subject: Pork and Papaya Salad
       Message-ID: <5BBP2SB@taronga.com>
       Date: Mon, 29 Mar 1993 06:51:47 GMT
       Lines: 23

       1/4 cup dried currants
       1/2 cup balsamic vinegar
       1/4 cup walnut oil
       1/4 cup chicken broth
       1 tablespoon honey
       1/4 teaspoon ground cinnamon
       1 pound cooked boneless pork loin roast
       1 head Belgian endive
       Bibb lettuce leaves
       2 papayas, seeded, peeled and sliced lengthwise
       2 avocados, seeded, peeled and sliced lengthwise
       1/4 cup broken walnut pieces

       In a small bowl pour enough boiling water over currants to cover.
       Let stand 5 minutes; drain. For dressing, in a screw-top jar
       combine vinegar, oil, chicken broth, honey, and cinnamon. Cover;
       shake well. Trim fat from pork; slice thinly. Separate leaves of
       Belgian endive. Line 6 salad plates with lettuce leaves. Arrange
       pork, endive, papaya, and avocado on plates. Sprinkle with
       currants and walnuts. Drizzle dressing over salads.

       Stephanie da Silva                            arielle@taronga.com

       -----------------------------------------------------------------

If you give more than one keyword, then all documents containing any of the keywords will be listed.

Learning more about WAIS

A bibliography of documents, services and sources for WAIS is maintained by Barbara Lincoln Brooks of WAIS Inc. The bibliography is available from ftp.wais.com in the directory /pub/wais-inc-doc along with many other WAIS documents. WAIS Inc. provides commercial WAIS software and services. The support for the freely available version, called freeWAIS, is assumed by the Clearinghouse for Network Information Discovery and Retrieval (CNIDR).

For information on freeWAIS software contact freewais@cnidr.org

There are currently four main FTP sites for WAIS documentation and software:

ftp.cnidr.org
ftp.wais.com
quake.think.com
sunsite.unc.edu

Mailing list: wais-discussion@wais.com To subscribe send a mail to wais-discussion-request@wais.com

Usenet newsgroup: comp.infosystems.wais

WAIS was developed at Thinking Machines Corporation.