"A Rant About Usenet"  is a passionate argument against "machine-generated spew" that is well worth reading. (It is the source of the famous quote "by God I KNOW what this network is for, and you can't have it".)
Whittaker, S., Terveen, L., Hill, W., and Cherny, L. (1998) The dynamics of mass communication. In Proceedings of the 1998 ACM conference on Computer supported cooperative work, Seattle, Washington, pp. 257-264.
[Hundreds of people have written thousands of pages about Usenet. The following was written before the author realized he was being redundant. ;-] --CliffordAdams
Originally UseNet was essentially controlled by the members of the "backbone cabal", which decided whether to carry newsgroups or not. Newsgroups could be created simply by a well-known person sending out a "control" message. Later, after some disputes, the "voting" system was introduced (and is still used today). The "alt." hierarchy remains as a "non-controlled" area, although most admins will at least limit newsgroup creation/removal messages to "reasonable" newsgroups.
The "alt." hierarchy:
The "alt." (alternative) hierarchy was originally a rather minor part of UseNet, until the great "rec.sex" vote (which I witnessed). Briefly, a proposal for "rec.sex" was made, and passed by a record number of voters. The backbone admins said "nope"--they weren't going to carry it. (At the time much of UseNet was carried over *very* expensive links that were officially dedicated for other uses.) The logical move was to use the "alt." hierarchy. Many admins found it much easier to carry all "alt." newsgroups rather than trying to justify any particular groups.
Today, the "alt." groups are often becoming fragmented, except for the few well-known groups. While anyone can create an "alt." newsgroup, few "unrecognized" creations will be effectively carried on most systems.
Some more history on rec.sex...The backbone cabal co-existed with the voting system for some time. When rec.sex was proposed, and voted on, it did indeed receive the record for UseNet voting. It was GeneSpafford? at GeorgiaTech? (where he worked as a systems' administrator) who objected to the creation of the group. I cannot recall any official reason for his decision every being made public: whether it was his personal opposition, or came from higher-ups at GeorgiaTech?, or what.
The backbone cabal refused to carry rec.sex on any of their systems. The next move was to propose the newsgroup comp.protocols.tcp-ip.eniac (ostensibly, to discuss the implementation of TCP/IP on that rather odd 17-bit architecture). The group was proposed as a way to prod the backbone cabal: to prove that they didn't care beans for votes. A vote was held. and comp.protocols.tcp-ip.eniac won a record number of votes (more than rec.sex), because many people had voted against rec.sex, but were appalled that their vote didn't apparently matter at all, and so they were happy to join the protest.
Now UseNet had always had the philosophy that you gain nothing by not creating a popular newsgroup: the discussion will just take place elsewhere. So when the cabal (predictably) refused to create comp.protocols.tcp-ip.eniac (on the newly-created grounds of frivolousness), people began discussing Eniac implementation of TCP/IP elsewhere in customary UseNet fashion.
This additional step was crucial, because it demonstrated that the cabal could not function adequately. Merely creating alt.* wasn't the point (beneficial though that may have been); the point also was to show that the cabal was basically not willing to abide by votes. The result ultimately was UseNet as we know it today: