Big-8 Usenet hierarchies

From Usenet Big-8 Management Board

The Big 8 is a collection of international discussion newsgroups covering a wide range of topics organized in eight hierarchies. Discussion is primarily in English, though there are exceptions, most notably in the soc.culture.* sub-hierarchy.

While international in scope, many newsgroups in the Big 8 have a United States flavor. This dual nature reflects the historical origins of the Big-8, and the relatively large share of native English speakers who live in the United States.

"A Usenet hierarchy is, reduced to its essence, a set of Usenet newsgroups that share a common naming prefix, such as all groups starting with 'comp.' or all groups starting with 'de.' The names of Usenet newsgroups define a hierarchy of names, with '.' used as the separator between the levels of the hierarchy ... The most significant part of the name is given first. The first component of the name is special and more significant than the rest of the name, since it defines the top-level Usenet hierarchy to which that group belongs" (Allbery).

The newsgroups that form the Big-8 hierarchies use the following naming prefixes:

comp.* Computer topics, both hardware and software.
news.* Administration of the Big 8, as well as about Usenet and Netnews in general, and related topics.
sci.* Science and technology.
humanities.* The humanities.
rec.* Recreational topics, including music, sports, games, outdoor recreation, hobbies, crafts, ...
soc.* Socializing, society, and social issues.
talk.* Endless discussion, largely about politics.
misc.* A mixture of newsgroups that don't fit the other 7 hierarchies. Many are about the practical aspects of everyday life.

The complete list of Big-8 newsgroups is in a special text file called "Checkgroups".

A note on the soc.* hierarchy:

On Tue, 11 Jul 2006 21:14:08 GMT, Jim Riley (jimrtex@pipeline.com) wrote in AwUsg.6099$PE1.2213@newsread2.news.pas.earthlink.net:

The soc.* hierarchy has always been mostly about socializing, rather
than society or social issues.  If you look at the groups placed in the
hierarchy as the result of the Great Renaming, soc.singles might well be
considered the archetype for the hierarchy.

Groups such as misc.legal and misc.kids (about parenting) were not
placed in the soc.* hierarchy because they were not seen as fitting the
hierarchy, even though they are important topics relating to how society
is organized.

See the charters for soc.atheism and talk.atheism for an example.
soc.atheism was designed as a place where non-theists could hang out
with out harassment from theists.  If they wanted to debate politics or
religion (or the lack thereof) they could use talk.atheism.

Nevertheless, the soc.* hierarchy does indeed include some groups that are about society and not about socializing. For example, soc.org.nonprofit is about the creation and operation of non-profit organizations.

History of the Big-8

"It is assumed in some quarters that if you missed the early days of
Usenet and didn't use Lynx from a Unix command line, you missed the
Magic of Internet Culture" (The Cluetrain Manifesto).
Usenet Wikipedia article on Usenet
Big_8_(Usenet) Wikipedia article on the Big-8
A Google search--news.groups & history
Joe Bernstein's Web Site
Archive for the History of Usenet Mailing List
How the Management of the Big-8 Developed

In the Beginning: "Let's Play 'Name that Froup!'"

(The 'froup' typo is attributed to Laura Zurawski in the alt.stupidity FAQ at MIT's RTFM site.)

On Sat, 18 Dec 2004 19:40:56 -0800, Richard Henry wrote: >I don't understand the logical foundation of the voting process, I guess.

Jim Riley replied:

Once upon a time groups were created simply by posting a message to the group. The news software would see an unrecognized name and create the group. New froups were created due to a misspelled name.

So a process was developed where a person who thought a group should be created would post a message to (example, "I'm thinking about creating a screenwriting group. Is there anyone interested?"). People would send the "proponent" e-mail expressing their interest, or perhaps expressing that they didn't think that the group was a good idea.

If there was sufficient interest in the new group, and not too much opposition, the proponent would post a message to the group, causing it to be created. A group that was created in accord with this process would be included in canonical lists of newsgroups, increasing the likelihood that it would be created on more, if not most, sites.

Over time, "sufficient interest" became quantified, and eventually became 100 more Yes votes than No votes. I don't know of any particular reason that the Yes and No votes are compared in the way that they are; it may be simply be a way to have No votes be registered with typically little consequence. At one time, most proposed groups received over 200 Yes votes, whether there were 10 or 30 No votes didn't matter.

At one point, there was talk of adding a 2/3 supermajority rule, but it was unclear whether this was to permit groups with margins such as 120:40 to be created or to stop groups that had a 500:300 margin from being created. The former group had sufficient interest, and not too much opposition. The latter group had a lot of opposition, despite achieving a 100 vote plurality.

A formula that included a supermajority could not be worked out, and the idea was dropped. About 6 months later, sci.aquaria was proposed. It was widely opposed because of its placement in the sci.* hierarchy rather than the rec.* hierarchy. The group passed by something like a 700:400 margin under questionable circumstances. Someone noticed that if there had been a 2/3 rule in place, the group would have failed, and the current 2/3 rule was added.

sci.aquaria was created under the old rules. Since that time, few groups have failed because of the 2/3 rule, since it is rare for a proposed group to receive 100 No votes, and even if there is a lot of No votes, there might be even more Yes votes.

At one time the average vote was 200+:30. Most proposed groups were created, and since they had sufficient interest they had some level of success. Groups with somewhat marginal interest might squeak by on a 135:30 vote.

A few groups would fail. If it was 20:30, the group probably would have been unsuccessful. If a group failed 120:30 it is not seen as a flaw in the system, but rather simply a disappointed proponent.

The proposals with massive opposition and massive support were rare, and had gone beyond the point of determining whether there was interest in a _newsgroup_ about Topic X, but were rather popularity contests aobut Topic X itself.

History of news.groups

  • In the beginning, all users were expected to read net.general. The purpose of the group was "initial queries and announcements." Proposals for and announcement of new groups took place in that group.
  • net.news.group was created to remove that traffic from net.general, according to Joe Bernstein's History:
net.news.group
  Proposed as net.names December 26, 1981 by Jerry Schwarz.  (He
    had proposed a net.groups, similar in purpose to the
    net.newsgroups that Curt Stephens later proposed, on June 14,
    1981!  Obviously a far-sighted person.)
  Proposed as net.news.groups January 13, 1982 by Jerry Schwarz.
  Announced January 15, 1982 by Mark Horton.
  Archived January 15, 1982 to ...
  Listed January 26, 1982 to November 1, 1986 without interruption.
  Description line (1/26/1982):  for discussions about proposed new
    newsgroups,
  Description line (11/1/1986):  Discussions and lists of newsgroups
    (Joe Bernstein's comment: Sound familiar, anyone?)
  • Two early posts described the purpose of net.news.group. The first post also shows the beginnings of the effort to maintain a list of newsgroups. The second shows that users could still create new newsgroups at will on some systems.
Message-ID: [bnews.azure.1484]
Newsgroups: net.general
From: curts@sri-unix
Date: Sat Nov 20 10:30:44 1982
Subject: grouplist

Some time ago, I regularly submitted a list of newsgroups to the net. Each 
entry to this list contained a brief summary of the purpose of that group. 
I had to stop, when I found myself too busy to keep up. Now, I am going to
try to maintain the list again.

...

A new newsgroup may be created by simply posting material to the 
net under a new newsgroup name. However, THIS IS NOT RECOMMENDED! 
There are limits to the number of newsgroups that can be supported
by the net. If you wish to send material to the net, first try to 
find an established newsgroup that deals with a subject related to
that material. If there is no appropriate newsgroup, suggest the 
creation of a new group via net.news.group. Usually, there will be 
enough feedback to establish whether there is an audience for the 
subject that you would like to discuss. 

...

--------------------------------------------------------------------------
 net.news.group      for discussions about proposed new newsgroups,
--------------------------------------------------------------------------

...

Curt Stephens

Tektronix 
Beaverton, Oregon, USA

--------------------------------------------------------

From: cbosgd!mark (cbosgd!mark)
Subject: official policy on creation of new newsgroups 
Newsgroups: net.news.group
Date: 1982-05-31 21:52:27 PST 
Message-ID: [bnews.cbosgd.2329]

The current official policy on creation new newsgroups is that someone
is supposed to propose it in net.general,net.news.group (substitute
whatever for net if it's a local group, or set up your own procedure) 
and all followups are supposed to go to net.news.group only.  (The followup
command in all recent versions of news does this automatically.)  Once
a consensus is reached, the person who originally proposed it should
make a yes/no decision on the newsgroup and a name, announce it to
net.general, and create it.  (If he is not the netnews contact at his
site, he should ask that person to create it, or ask me.)

Modern versions of the software do not let people create new groups
unless they are the local netnews contact, or the super user.  There
are many old versions still in use out there that don't enforce this,
which shows that no matter what change we make, getting software to
enforce it in the finite future is very hard.

The net.sctv business has recreated the suggestion of a netnews
committee who makes all the decisions.  It has not been resolved yet.
The last time somebody suggested this, there was overwhelming sentiment
to not restrict new newsgroups, but rather to make the software wonderful,
by such things as automatically expiring empty newsgroups, and making it
much faster to do readnews or checknews.  (This will probably happen in
version 3.0, which is right now in the wishful thinking stage due to
lack of time on the part of the authors.)

As to natural language, the flaming has gone on for a while but the person
who proposed it hasn't made a decision.  Personally, I don't understand
what you guys are going to talk about, but there is lots of stuff I don't
subscribe to.  So I'll make a decision:  net.nlang is now created with no
subgroups.  Post to it and mention the particular language in your title.
Any language that picks up significant activity that continues after the
novelty wears off (the participants should decide this) should create its
own subgroup.  New natural language newsgroups can be started in the parent
newsgroup.  (Off the wall thought - once the 14 char limit goes away
in the eventual version 3.0, you might want to call it "net.tongues".)

        Mark 
  • By 1982 (less than a year after its creation), net.news.group was nominated as "the silliest group on the net" and the first cries were heard to "go back to an earlier time" (!):
Message-ID: [bnews.unc.3498]
Newsgroups: net.news.group
Path: utzoo!decvax!duke!mcnc!unc!wm
X-Path: utzoo!decvax!duke!mcnc!unc!wm
From: unc!wm
Date: Sat May 29 07:12:46 1982
Subject: DOWN with net.news.group
Posted: Fri May 28 15:51:51 1982
Received: Sat May 29 07:12:46 1982

Personally I think half the news groups that have been started
are silly.  Natural language groups fit into this category, but
WHO CARES?  Anyone can start a new group, and there are several
new groups started every week or so by someone who doesn't know
the "rules" yet.  The only purpose net.news.group seems to serve
is as a forum for people to flame about how other people's ideas
for new groups are silly, or how something should be a mailing
list, or should be in net.misc, or shouldn't be in net.misc, etc.
I therefore nominate as the silliest group on the net, you guessed
it, net.news.group!  Let us go back to an earlier time, when
all you did if you thought there was an interest in a new group
was to post an article to it.  If nobody responds, or if the
group sees activity for two weeks and then dies, it should be
deleted by some automatic deletion mechanism.  What are the chances
of seeing this installed in the near future?  Of course, we will
still see long discussions about what the name of a group should
be (you know, I think I will go back into the archives and see
what percentage of articles in a group are related to what its
name should be.  Any bets?).
				unc!wm
  • In 1986, net.news.group became news.groups:
From: mark@cbosgd.ATT.COM (Mark Horton)
Newsgroups: net.news.group,news.groups
Subject: net.news.group is being renamed news.groups
Message-ID: [3039@cbosgd.ATT.COM]
Date: Fri, 7-Nov-86 15:14:38 EST
Article-I.D.: cbosgd.3039
Organization: AT&T Bell Laboratories, Columbus, Oh
 
This newsgroup is being renamed from net.news.group to news.groups.
This renaming will gradually take place over the next few weeks.
More and more messages posted to this newsgroup will be aliased
into the new newsgroup as they pass through the net, and people
will begin to post to the new group.  After a few weeks, the
old name will be removed.
 
This note is to inform you of the renaming so you can begin to
read the new group as well as the old group.
 
 	Mark Horton
 	Director, the UUCP Project
  • A comment from Jim Riley on the early history:
Message-ID: <cd0xg.12471$ug4.4262@newsread2.news.pas.earthlink.net>
Date: Wed, 11 Oct 2006 06:32:02 GMT

...

The name of news.groups
is more or less an accident.  net.news.group was chosen because it fit
in 14 characters, and the last 's' was added during the great
renaming.

...

Spaf

Spaf is the nickname of Eugene H. Spafford, who is a professor at Purdue University who specializes in issues of information security, digital forensics, privacy technology, and security and science policy. Among many other things, he is the founder and executive director of CERIAS, and chair of ACM's US Public Policy Council, USACM.

  • Published the first set of newusers documents for Usenet ca. 1982 (and continued to maintain and post them for the next 11 years).
  • Developed the concept of the Usenet backbone ca. 1983.
  • Created the idea of the Usenet newgroup "ballot" ca. 1987.
  • Was moderator of one of the first moderated newsgroups not represented by a gated mailing list (news.newusers).

The Failed Trial Group System (1991)

Brad Templeton explains the system.

tale's Reform Proposal (1993)

Available in ftp://ftp.isc.org/pub/usenet/news.announce.newgroups/comp/comp.dcom.telecom.tech.

From tale@uunet.uu.net Thu Oct 14 20:50:30 1993
Path: uunet!tale
From: tale@uunet.uu.net (David C Lawrence)
Newsgroups: news.announce.newgroups,news.groups,news.admin.policy,news.admin.misc
Subject: comp.dcom.telecom.tech and Changing the Guidelines
Followup-To: news.groups,news.admin.policy
Date: 14 Oct 1993 19:01:27 -0400
Organization: UUNET Communications
Lines: 78
Sender: tale@rodan.UU.NET
Approved: tale@uunet.uu.net
Message-ID: 29klo7INN5sq@rodan.UU.NET
References: 29esvpINN4ae@rodan.UU.NET
NNTP-Posting-Host: rodan.uu.net
Xref: uunet news.announce.newgroups:4160 news.groups:84553 news.admin.policy:7211 news.admin.misc:5076

This message is intended to address my recent suggestion that the
comp.dcom.telecom.tech group be created despite its failing vote result,
and to give an introduction toward ideas I have had for a while
regarding restructuring the Guidelines for USENET Group Creation.  I
would like to keep it relatively short, and some people might believe
there are glaring omissions.  I hope that all points will be addressed
in the ensuing discussion.

First, I thank everyone who has written to me over the past two days.
Nearly all of the messages that I received were carefully written and
well presented.  I'm sorry that I cannot reply individually to all of
them.

The diversity of the mail has been surprising at times.  I've got
messages from people who voted for the proposal asking me not to create
it, and messages from those who voted against it remarking that I should
create it.  I've got praise from people who are happy to see the spirit
of the Guidelines upheld, and derision from those who would hold me to
the letter of them.  I've got more than ample indication that if I did
send the newgroup without further debate that it would be widely
honoured and the group well propagated.

Before I continue, I want to make clear my primary source of bias
regarding USENET group creation.  I support the creation of well focused
newsgroups which are not duplicitous.  I have yet to oppose a USENET
newsgroup based on its content.

Voting by the hundreds is not a good way to administer a classification
system.  "No" votes are not particularly effective at the job they were
intended to do, which was to stop poorly formed proposals.  They have
been used to stop groups which should have been created and have not
stopped those which shouldn't.  I don't believe that any manner of
playing with the numbers and ratios will resolve this inequity.  They
should be abolished.

Namespace management is best served by a small working group who can
help ensure consistency and coherency.  The net supported this idea two
years ago when a proposal to use a "trial" process to create groups was
approved by a traditional USENET vote.  One of its features was that a
small group of people were to chose the name of the group.

I would like to advance a proposal for new guidelines for group
creation.  The essential differences would be to make new group creation
easier and shorter for people interested in the topics to be discussed,
at the cost of not being able to push any name and ill-formed charter
forward.

I will be working on the draft of these Guidelines and submitting them
for discussion in the near future.  Please do not start debating these
ideas until you see how they come out in the proposal.

Allbery's Reform Proposal (1996)

From: Russ Allbery <rra@cs.stanford.edu>
Subject: [DRAFT] RFD: news.groups reorganization
Date: 1996/10/24
Message-ID: qumramogu12.thoron@cyclone.stanford.edu
X-Deja-AN: 191793806
sender: eagle@cyclone.stanford.edu
organization: The Eyrie
newsgroups: news.groups


[ This is a draft only.  Comments are very much welcome.  Anyone else
  who is considering a reorganization of news.groups, please contact
  me; I would greatly prefer to work together rather than end up with
  competing proposals. ]

From: Russ Allbery <rra@cs.stanford.edu>
Subject: RFD: news.groups reorganization
Newsgroups: news.announce.newgroups,news.groups

		     REQUEST FOR DISCUSSION (RFD)
		 moderated group news.groups.creation
              moderated group news.groups.reorganization
               moderated group news.groups.innovations
             unmoderated group news.groups.pre-proposals
      unmoderated group news.groups.policy (renames news.groups)

Newsgroup lines:
news.groups.creation		Creation of single newsgroups.  (Moderated)
news.groups.reorganization	Reorganization of newsgroups.  (Moderated)
news.groups.innovations		Innovative newsgroup concepts.  (Moderated)
news.groups.pre-proposals	Preliminary newsgroup proposals.
news.groups.policy		Policy issues of newsgroup creation.

This is a preliminary Request For Discussion (RFD) for the
reorganization of news.groups.  This is not a formal RFD.  Procedural
details are below.

RATIONALE: all groups

news.groups is becoming increasingly difficult to follow because of
the noise, poor use of subject lines, and overload due to multiple
group discussions happening simultaneously.  This proposal is designed
to break up that discussion and mark the actual group discussion so
that it can be easily filtered and followed by people only interested
in one proposal.

The core of the proposal is robomoderation of the two main groups to
require that every post be marked with a tag specific to a particular
proposal. 

...

fa.*

On Fri, 20 Oct 2006 10:20:52 +0200, Kjetil Torgrim Homme <kjetilho@kaksi.ifi.uio.no> wrote in <1rbqo7iejf.fsf@buri.ifi.uio.no>:

>   > I run the gateway at fa.*, including fa.haskell.  I'm sorry to
>   > say I haven't put much effort into maintaining subscriptions the
>   > last few years, and as you allude to, many of the groups in the
>   > hierarchy are dead.
>   
>   Are you saying that the gated mailing lists themselves are dead?
>   That's too bad.

a little of both, I think.  (I know fa.haskell is alive, though.)

>   > all the fa.* groups are one-way, btw.
>   
>   I didn't know that. Was this always the case, or a deliberate
>   attempt to keep Usenet spam out of Mail?

our gateway has always been one-way.

a little background: as you probably know, the FA name is from before
the Great Renaming (it stands for From ARPAnet).  I don't know how it
worked back then.  some years later, Anders Ellefsrud, my predecessor
as news admin here at ifi.uio.no, thought that it was wasteful that
our students subscribed to mailing lists individually, and set up a
gateway to save bandwidth and to get automatic expiration.  he chose
to reuse the name fa.* which was no longer in use on Usenet.  this was
a local hierarchy at first, but some neighbours thought it was useful
and wanted feeds, and in 1998 it started leaking to the Usenet at
large.  this was under my watch, but I thought "hey, indexing at
Google is actually pretty neat", so I didn't try to find the leak.