In private email, a regular contributor here asked me about newsgroups, which have been mentioned on SAS-L but not explained. I decided to send an expanded version of my explanation to the list as well.

In short, a newsgroup is similar to a bulletin board or AOL forum. People (usually called "posters") submit (or "post") messages, and other people can read and respond to them. Messages are organized so that similar messages are presented together.

This sounds like a mailing list, and indeed newsgroups and mailing lists often serve the same purposes. The main differences between the two are how messages are distributed and how messages are read.

A mailing list, such as SAS-L, runs everything through a single server (you can also think of it as a single point of failure). If I want to send a message to everyone on SAS-L, I just mail it to sas-l@listserv.uga.edu, and it gets distributed to a centrally maintained list of subscribers. (SAS-L is a peered mailing list, meaning that there are actually several servers, but that doesn't change the workings in any essential way.)

With a newsgroup, there's no central point of control and no single list of subscribers. Every site that carries news (a much smaller number than the number of sites carrying mail) has software called a news server. If I post a message to comp.soft-sys.sas, which is the mirrored version of SAS-L, my news server assigns it a unique ID number, which usually contains the host name and some pseudo-random characters. Suppose the ID is 12345.newsguy.com. Later in the day, my news server will connect to neighboring news servers with which it has agreed to share news. It will ask each server in turn "Do you have message 12345.newsguy.com?" If the answer is no, it will send my message to that server. When that server, in turn, connects to its neighbors, the process will be repeated, and eventually my message will reach every server in the world that carries comp.soft-sys.sas.

In the old days, this process might take a long time - maybe a week to get to Europe, and much longer to get to remote parts of the globe (the connections were originally done over dial-up lines at night). Now it's very fast -- major news servers are constantly connecting to each other over high-speed internet lines.

To read and post news, you use a news reader. Common news readers are Netscape Messenger, Forte Agent (what I use at home), and Microsoft Outlook. On Unix systems, many people use trn or xrn. Just how you select newsgroups, read messages, and post varies between readers. The reading process is somewhat similar to the distribution process -- your news reader keeps track of the articles it's read, and when you connect to the news server it uses that list to figure out which messages are new and should be displayed. [1]

By the way, there's another kind of newsgroup that does have a central point of control. Moderated newsgroups, often used for controversial subjects, require postings to have moderator approval before being distributed. When you post to a moderated newsgroup, the message is mailed to the moderator, who approves it and injects it into the news stream from his or her own machine.

There's a bidirectional mail-to-news gateway between SAS-L and comp.soft-sys.sas. In theory, everything which appears in one will eventually appear in the other. For various obscure reasons, this occasionally breaks down, but one good feature of the gateway is that it lets us read and post via email, but use the newsgroup search features at Google.

You can read, search, and post newsgroup messages through the web using Google Groups -- click on the Groups tab on the main page. I don't use that interface for reading, because it's too hard to keep track of what I've already read, but I do use it for searching, and occasionally for posting. For some people, doing everything through the web is more convenient.

There are more than 40,000 newsgroups.

[1] This is an oversimplification. For one thing, the news reader and news server use a short, locally assigned message number rather than a message ID, to save space and make processing easier.

-- JackHamilton@FirstHealth.com Manager, Technical Development Metrics Department, First Health West Sacramento, California USA