Welcome to WAIS

            MXQWAIS -- A Motif WAIS Client
           (Wide Area Information Servers)
              Version 1.0, Oct. 8, 1992
          Erik L. Scott, escott@eos.ncsu.edu

This program is still being developed. Click on help
every few days and see what's new!
For help, click on the topic you want to know more about.
***END
Beginning WAIS
To see how WAIS works, try the following.
Click on the "Add to Server List" and wait for a list of
"sources" to pop up.  Select the "NCSU_FacStaff.src" source,
and click on "OK".  The NCSU_FacStaff.src source is listed
in the "By asking these servers..." list.  Now, in the text
entry box below "Tell me about...", type "Dick Sheridan".
Click on "Search" at the bottom of the window.  A window
will pop up in a moment with a list of names, including
"Sheridan, Dick" near the top.  Click on this entry, and
then press the "View" button at the bottom of the new
window.  In a second or two, a new window with information
about him will be displayed.  The "Search For Keyword" button
searches for words you used in your search.  The "Save a
Copy" button lets you save this window's contents to a file,
and the "Close" button will close this window.

This is just a "get your feet wet" example.  Read the rest
of the help, and by all means experiment with the software. 
You can't hurt anything by playing, and you may find some
things that are really useful.
***END
What is WAIS?
For the definitive answer on what WAIS is, here is a copy of
"Overview of Wide Area Information Servers" by Brewster
Kahle.  He invented WAIS, so I think it makes sense to use
his overview:

       Overview of Wide Area Information Servers
                    Brewster Kahle
                      April 1991


The Wide Area Information Servers system is a set of
products supplied by different vendors to help end users
find and retrieve information over networks.  Thinking
Machines, Apple Computer, and Dow Jones initially
implemented such a system for use by business executives. 
These products are becoming more widely available from
various companies. 

What does WAIS do?

Users on different platforms can access personal, company,
and published information from one interface.  The
information can be anything: text, pictures, voice, or
formatted documents.  Since a single computer-to-computer
protocol is used, information can be stored anywhere on
different types of machines.  Anyone can use this system
since it uses natural language questions to find relevant
documents.  Relevant documents can be fed back to a server
to refine the search.  This avoids complicated query
languages and vendor-specific systems.  Successful searches
can be automatically run to alert the user when new
information becomes available. 


How does WAIS work?

The servers take a user's question and do their best to find
relevant documents.  The servers, at this point, do not
"understand" the user's English language question; rather
they try to find documents that contain those words and
phrases and rank them based on heuristics.  The user
interfaces (clients) talk to the servers using an extension
to a standard protocol Z39.50.  Using a public standard
allows vendors to compete with each other while bypassing
the usual proprietary protocol period that slows
development.  Thinking Machines is giving away an
implementation of this standard to help vendors develop
clients and servers. 

What WAIS servers exist?

Even though the system is very new, there are already
several servers: 

  * Dow Jones is putting a server on their own DowVision
network. This server contains the Wall Street Journal,
Barons, and 450 magazines.  This is a for-pay server. 

  * Thinking Machines operates a Connection Machine on the
Internet for free use.  The databases it supports are some
patents, a collection of molecular biology abstracts, a
cookbook, and the CIA World Factbook. 

  * MIT supports a poetry server with a great deal of
classical and modern poetry.  Cosmic is serving
descriptions of government software packages.  The Library
of Congress has plans to make their catalog available on
the protocol. 

  * Weather maps and forecasts are made available by
Thinking Machines as a repackaging of existing information.

  * The "directory of servers" facility is operated by
Thinking Machines so that new servers can be easily
registered as either for-pay or for-free servers and users
can find out about these services. 

How can I find out more about WAIS?

        Contact Brewster Kahle for more information on the
WAIS project, the Connection Machine WAIS system, or the
free Mac, Unix Server, and X Window System interfaces. 
There is a mailing list that has weekly postings on
progress and new releases.  To subscribe, send an email note
to wais-discussion-request@think.com. 

Brewster Kahle
Project Leader
Wide Area Information Servers
Brewster@Think.com

***END
Finding New Sources
When you run mxqwais for the first time, a small list of
sources is presented to you.  This list is the same for
every user.  It's anything but a complete list of the
sources available.  To find other sources, ask the directory
of servers about them.  Use the "Add to Server List" button
to select the source "directory-of-servers.src".  Then in
the "Tell me about..." field, type in a few words or a
phrase about what you'd like to search for.  A good example
would be "weather".  Search for it.  In a few seconds, a
list will pop up with "weather.src" in it.  Select that line
and click on View.  After a pause of maybe as much as a
minute, a large window will pop up with a lot of meaningless
text in it.  It isn't really meaningless, of course, but
most people will only be interested in the large text field
at the bottom, the "description".  Read over the
description, and then select the "Save" menu command on the
File menu.  Then select the "Exit" menu command on the
File menu.  Now when you "Add to Server List", you'll have a
new selection available: weather.src.  Good questions to ask
this source include "Raleigh" and "GIF".  The latter is an
acronymn for "Graphics Interchange Format", and is how
weather maps are distributed.  Search for "GIF", and look
for returned lines in the form of usa-600x300-CIMMDDHH.GIF
and usa-600x300-CVMMDDHH.GIF.  The "CI" lines are infrared
spectrum maps, and the "CV"s are visible light images.  The
MM is a pair of numbers for month, the DD means date, and
the HH is hour (in GMT!).  So usa-600x300-CI100813.GIF is an
infrared spectrum map for Oct. 8th at 10:00am.  Select an
appropriate map and click on "View".  In a few moments, the
weather map will be displayed (it looks okay on mono
machines, but it's just stunning on a color machine.  Far
better than TV.)  When you're finished looking at the image,
press the spacebar in that window.

The process outlined above is how you go about discovering
new sources.  Start by asking the directory of servers about
what you want to know (vague terms work pretty well; for
example "NSF" will work a lot better than "principal
investigators who drive Volvos" if you're looking for NSF
grant information).
***END
Query Hints
Writing a good query isn't hard at all, but there are some
subtleties to consider.  WAIS looks through a collection of
files and returns the "headlines" for each file based on how
many times your search words occur in the file.  So if you
search for "automobile", the file that contains the word
automobile the most times is returned first.  The problem
begins when more than one search word is specified. 
Consider searching for "automibile brake lights".  Instead
of getting files that contain the phrase "automobile brake
lights", you'll get every file that contains "automobile",
every file that contains "brake", and and every file that
contains "light" or "lights" (WAIS thoughtfully ignores
trailing 's'es).  In a perfect world, the files that contain
the entire phrase "automobile brake lights" will be near the
top; but if there was a file that contained "light" a few
dozen times, then that file will probably be on top.  There
isn't much you can do about it, really, except to be aware
of it and try shorter query strings first.
***END
More to Come...
There will be new features added to the mxqwais client, as
well as to the servers on the Internet that provide the
information.  It's a good idea to keep an eye on this space
from time to time and see what's new.

To learn more about WAIS, see the following files:

eos% add wais
eos% cd /ncsu/wais/doc
eos% ls

If you have any questions, problems, or comments, let me
know.  My mail address is escott@eos.ncsu.edu.  Your
feedback helps me improve the program I write, and your
questions let me know what needs to be made clearer.
***END
