Path: utzoo!attcan!utgpu!watmath!mks.com!egisin
From: egisin@mks.com (Eric Gisin)
Newsgroups: news.admin
Subject: News Versions, another approach
Message-ID: <1989Nov15.213552.9499@mks.com>
Date: 15 Nov 89 21:35:52 GMT
Organization: Mortice Kern Systems Inc., Waterloo, Ontario, CANADA
Lines: 28

I wrote a short awk program that identifies news versions
based on Message-ID syntax. It wasn't too accurate
because B news, Notes, VMS, and P news all use the same syntax.
I did discover that there are several other versions
that I could not identify, which I have called E, and F and Other.

This analysis differs from the version-control-message method in two ways.
First, it only recognizes sites that have posted in the last two weeks.
Second, it also recognizes sites that do not respond to version messages,
which makes this method more reliable. Many C news sites don't seem to respond.
It is also including many mailing list gateways.

Here are the results from running it on a 2.4MB history file (2 weeks).

type	# distinct domains
B	2162
C	477
E	36
F	448
Other	325
Total	3448

Here are the REs I used in the awk program. ID is <$2@$3>.
$2 ~ /^[0-9]+$/	{		# B news
$2 ~ /^(1989|[A-Z0-9]*\.89)[JFMASOND]/ {		# C news
$2 ~ /^[JFMASOND].*1989/ {	# E news
$2 ~ /^[0-9.]+AA/ {		# F news
{				# Other