Title: Release Notes for freeWAIS 0.1
Authors: Jim Fullton, Kevin Gamiel, Jane Smith, Robin Braun
Copyright: Copyright 1993 CNIDR
Last Updated: 04.09.93
Abstract: This file contains a list of fixes and enhancements
for freeWAIS Version 0.1.

General Notice:

Copyright notices have been included for the developers
who wished CNIDR to do so. CNIDR has also added a compilation
copyright to the entire release to protect the work of the network
community. See the COPYRIGHT file for more information.


Specific enhancements to the waisserver/waisindex:

--
1.Compressed files are now supported by the server.

The server will uncompress files prior to indexing then discard
the uncompressed files afterwards. Server based retrieval of
compressed files has also been added.

Jonathon Goldman (jonathon@think.com) of Thinking Machines Corporation

--
2. A new scoring mechanism is introduced for weighting
returned items. 

Documents are now scored based on the log of the query word
frequency, the number of occurrences of the word throughout the
entire database, and the size of the given document.

Simon Spero (ses@sunsite.unc.edu) at UNC-CH

--
3. Synonym support is now included on a site specific level.

The server administrator can manually create a file called
database_name.syn containing a list in the form:

    word synonym synonym synonym synonym
    word synonym synonym synonym synonym
    word synonym synonym synonym synonym
    word synonym synonym synonym synonym

Any word on a line will match any other word on the same line.

        Example:

        heart myocardium myocardial cardiac

A document with the word cardiac will be returned as a hit for
a query containing the word heart. The synonym list is at
the administrator's discretion.  Note - when a synonym list is added
or changed, the database must be re-indexed.

Ahmet Taylor (Ahmet_Taylor@unc.edu) at UNC-CH

--
4. The symbol UNCOMPRESSOR has now been defined so that you can
choose which external application is used for decompression.

This is located in the futil.c file.

Francois Schiettecatte (francois@welchgate.welch.jhu.edu) at the Welch
Medical Library

Jonathan I. Kamens (jik@athena.mit.edu) of Aktis, Inc.

--
5. A headline cleaner has been added to remove carriage
return and line feed characters from the headline.

This is located in irtfiles.c and is called cleanHeadline.

Francois Schiettecatte (francois@welchgate.welch.jhu.edu) at the Welch
Medical Library

--
6. Support for stemming is included in this release.

The server stems the query words prior to searching. The stemmer
is contained in stem.c and can optionally be replaced.  The
default stemmer is a Porter stemmer.

Francois Schiettecatte (francois@welchgate.welch.jhu.edu) at the Welch
Medical Library

Jim Fullton (Jim.Fullton@cnidr.org) of the CNIDR

--
7. There is now support for partial word searching and optional
stop-list file support.

Don Gilbert (gilbert@bio.indiana.edu) at Indiana University
(see iubio-wais.readme and iubio-wais.news)

--
8. The ability to perform boolean and literal searches has been added.

Tim Gauslin of the USGS
(see iubio-wais.readme and iubio-wais.news)

--
9. Support is now in place for extra keywords and extra keyword
file indexing.

David Post (post@cs.cornell.edu) of the Design Research Institute

--
10. Support has been added for indexing and retrieving multitype
documents.

Francois Schiettecatte (francois@welchgate.welch.jhu.edu)

--
11. More granularity is now included in the -mem option. -mem
will now understand 2 through 5.

Jonathan I. Kamens (jik@athena.mit.edu) of Atkis, Inc.

--
12. Various fixes have been added which will allow wais to be built
more easily on different platforms.

Jonathan I. Kamens (jik@athena.mit.edu) of Atkis, Inc.

--
13. There is now a DUMPCORE compiler option which will force the
waisserver to dump core when aborting.

Jonathan I. Kamens (jik@athena.mit.edu) of Atkis, Inc.

--
14. The intermediate index files in the indexing process by waisindex
are now merged during the process rather than at the end.

This is more economical in terms of disk space. The END_MERGE
compile option will override this feature and merge the files at
the end of the indexing process which is how it used to work.

Jonathan I. Kamens (jik@athena.mit.edu) of Atkis, Inc.

--
15. The quality of the source descriptions has been improved by
automatically adding the fifty most frequently occurring words in the
index to the .src file.

AT&T and the Internic Team

--
16. Host and daatabase level server security has been added.

Francois Schiettecatte (francois@welchgate.welch.jhu.edu)


==============================================================================

Much thanks goes to all of the contributors to this release. Your
time and code are an invaluable part of freeWAIS-0.1 . We would
like to particularly thank Francois Schiettecatte at Johns
Hopkins University for all of his efforts in putting together
freeWAIS-0.1. CNIDR looks forward to continued cooperation and
work with all of you in the future.

If you or your site is interested in becoming an official beta
tester for CNIDR for future releases, please send your name and
email address to freeWAIS@cnidr.org. This will help us to have a
more organized distribution and test audience in the future.
Thanks.
