$Header: README,v 1.0 93/08/31 11:49:03 wjm Exp $

$Log:	README,v $
# Revision 1.0  93/08/31  11:49:03  11:49:03  wjm (Bill Middleton)
# Initial revision
# 

  This is the RegEx News package.  It can be used to scan USENET news for
Regular Expressions, and simple keywords.  It is designed to run for 
multiple users, either local or remote.  Each user must have a file in
their subscriber directory called .newsearchrc.  In this file are lines
of the following form:


immunology and pain
(hamster) and (duct\s+tape)
carasso
\shp\s and \bppp\b and not comp\.sys\.hp
spatial\s+((analysis)|(distribution))
routing and arp
\sgopher\s and not comp\.infosystems\.gopher
\bperl\b and not comp\.lang\.perl


  At startup time, the NewsScan script reads each of these lines from the 
individual subscriber's .newsearchrc's, and then scans the batches for each
one, article by article.  NewsScan also works by scanning articles out of 
the spool after they are processed, depending on which version you use.

  When a match is found in a given article, RegEx News prints out a line
to another file in the subscribers directory, called .hits.  This file is
later loaded by the reader script for each subscriber, to read the articles 
of interest.  If a .forward file exists in the subscriber directory, then
the hit is is mailed to the subscriber, along with any others during that
scan.  The remote subscriber then uses the reader script, and either his/her
local newsspool or an nntp connection, to read the articles.









