alt.humor.best-of-usenet

Long-form reference · 1,860 words · about an 8-minute read

The group and its era

Most humour newsgroups worked like an open microphone: you wrote your joke, posted it, and took your chances with the room. alt.humor.best-of-usenet inverted the model. Nothing was written for it at all. Its readers acted as scouts, forwarding posts they had stumbled across elsewhere on the network, and the moderators decided which finds deserved a wider audience. What appeared was humour that had already happened somewhere else — remarks made in passing, in the middle of earnest arguments, in groups devoted to anything but comedy.

The machinery was unusually tidy for the alt.* hierarchy. Submissions went to one address and correspondence with the moderators to another; each accepted item went out with a line crediting the moderator who had passed it, alongside the original author and the group the post had been lifted from. The group's FAQ listed its moderation contact at a Duke University address — an institutional fixed point of a kind most alt.* groups never had.

Moderation on Usenet was a postal arrangement rather than a technical privilege. An article aimed at a moderated group was not passed to the network at all: the posting software mailed it to the moderation address, and it reached the world only when it came back carrying an Approved: header line. A curated group thus rested on one line of a message header and on whoever held the mailbox behind it, and readers saw what came out but never what had gone in.

Its fortunes tracked Usenet's own: liveliest through the 1990s, when the network was still the internet's principal public conversation, and quieter once the web drew the crowds away. The FAQ, though, was still being maintained well into the 2010s — curation, it turns out, breeds persistence.

What was discussed

Strictly speaking, nothing was discussed here at all. The group was moderated and carried only the reposts themselves; conversation about them — including the durable genre of complaints about what the moderators had seen fit to let in — was directed to a companion group, alt.humor.best-of-usenet.d, which existed for precisely that overflow.

The staple was the accidental gem: the withering one-line reply that ended a flame war, the deadpan answer to a perfectly serious technical question, the earnest post whose author plainly had no idea how funny it was, the thread that drifted by individually reasonable steps somewhere ridiculous. Since the whole network was fair game, the group also read as a tour of Usenet's comic registers — dry pedantry from the technical groups, in-jokes from the hobbyist ones, and the alt.* hierarchy's dependable absurdism.

The rules put responsibility for a submission squarely on the person who forwarded it, and attribution was the point of the exercise rather than an afterthought. In an era when e-mail was filling with tenth-generation anonymous forwards, a repost that arrived with the author's name, the source group, and the selector's signature attached was a small act of citation — bibliography as a form of respect.

Because nothing originated here, the contents recorded two things at once: what Usenet had said, and what a few readers happened to notice. A repost is evidence of the group it came from and of the taste of whoever carried it out, and the two cannot afterwards be pulled apart.

A group made of other groups' postings

The best-of group was a small genre of its own on Usenet: a group with no native content, whose whole output was other groups' output moved one hop sideways. Nothing was written; something was chosen. That single decision creates two problems an ordinary newsgroup never faces — selection, meaning who decides and by what route anything reaches them to be decided about; and permission, meaning whether the author has any say in being published again elsewhere.

Selection in a repost group runs on volunteers noticing. No moderator can read the whole network, so the queue is filled by whichever readers happen to see something, remember the submissions address and forward it. The result is a sampling bias no editorial care removes: a best-of group covers the groups its scouts read and is silent about the rest of Usenet. The judgement that follows is a second filter, private and unexplained — which is why such groups so often acquired a companion .d group, the room in which a reader who thought the wrong things had been let through could say so.

The permission question was live because Usenet postings are ordinarily copyrighted works. Under the Berne Convention, protection attaches automatically once a work is fixed, with no registration or notice needed; the United States acceded in 1988 and the convention took effect there on 1 March 1989, ending the older American rule under which publishing without a notice could forfeit protection. Posting an article to a public network was not, on that reading, a gift of it. The counter-argument was that Usenet functions only by copying, so to post is to license redistribution; where that licence stopped — at the next news server, at a repost, at a printed collection — nobody could say authoritatively, and the question was argued in the humour groups for years without being settled there.

Editors drew the line in different places. rec.humor.funny's submission rules told readers plainly not to send in other people's postings: if you saw something funny in a group that had not been trying to be funny, you were to write to its author and encourage them to submit it, the only exception being postings made in the United States before April 1989 without a copyright notice. alt.humor.best-of-usenet was built the other way about, on the premise that a third party could carry a posting across so long as the author's name went with it. Two groups doing much the same job, and between their rulebooks the whole argument.

The humour groups around it

None of this happened in isolation. The rec.* hierarchy carried rec.humor, the unmoderated joke group and the volume end of the trade. Its own FAQ set out the life cycle of a joke there with some feeling: posted once, read by thousands, then reposted by the fraction of them who had not noticed the previous ten repetitions. Repetition rather than scarcity was the problem the moderated groups existed to solve.

rec.humor.funny was the celebrated answer to it — a moderated selection, running from 1987, and by the Usenet readership statistics Brian Reid compiled the most widely read publication on the network from 1989 into the mid-1990s, at an estimated 440,000 readers. Its published policy was to judge submissions on comedic structure rather than on content, and to encode the jokes it classified as discriminatory in ROT13, the trivial letter-shifting cipher Usenet used for punchlines, spoilers and anything a reader might want to decide about before seeing. An editorial policy expressed as a text transformation.

Beside the edited groups sat institutions that were participatory instead. The Usenet Oracle, renamed the Internet Oracle in 1996, ran on a question-and-answer conceit in which everyone who asked a question was sent somebody else's to answer, routed anonymously through a central program and published as a digest in rec.humor.oracle, with rec.humor.oracle.d for the argument about it. It was set up in 1989 by Steve Kinzler, a system administrator at Indiana University, from a version Lars Huttar had posted that year. Next door, alt.folklore.urban worked the seam where the funny story and the false story meet; the fact-checking site later known as Snopes grew out of its founders' work there.

Who decides what a site carries

Moderated humour became a governance subject in the winter of 1988-89, when an ethnic joke published in rec.humor.funny — the sort of item the group's own policy would normally have encoded in ROT13, as its moderator afterwards acknowledged it should have been — became the subject of a complaint that began elsewhere on the network and travelled from campus to campus. At Stanford University the central computing administration stopped carrying the group. The argument that followed was about who decides what a site carries: whether the people running an institution's computers are publishers, answerable for what sits on the disks, or carriers, obliged to pass on what the network sends.

The fullest account was kept by John McCarthy, then professor of computer science at Stanford. By his record the computer science department opposed the removal and went on carrying the group on its own machines; the Academic Senate's steering committee referred the matter to its Committee on Libraries, which argued that electronic media should not be treated differently from print, the test being cost and universality rather than content. The group was restored. Nothing in it bound anyone else, but it is where the question of who may drop a newsgroup stopped being administrative and became public.

Context

The obvious comparison is rec.humor.funny, the moderated joke group Brad Templeton founded in 1987, for a time among the most widely read things on the network. The two were complements rather than rivals: rec.humor.funny published jokes written to be jokes and submitted for judgement, while alt.humor.best-of-usenet collected the comedy Usenet produced without meaning to. One was a stage, the other a scrapbook.

It was also routinely confused with alt.best.of.internet, which gathered the funny material then circulating by e-mail and forwarding across the wider internet. The rec.humor.funny site's own guide to the humour groups acknowledged the muddle and drew the line that mattered: this group took Usenet postings only, and it was moderated. The distinction was not pedantry — the moderation queue was precisely what kept the group readable.

Humour also turned out to be among the best-preserved of Usenet's output. Curated material is easy to mirror: already chosen, already formatted, small enough to keep. rec.humor.funny's collection is still published on its own site, the humour groups' periodic postings survive in the Usenet FAQ archives, and much of the traffic itself still sits in the Google Groups archive, which stopped carrying new Usenet postings in 2024 but kept what it already held. That is why a reader today generally reaches a group like this one from a FAQ mirror rather than a news server: the index lasted better than the conversation.

The format has aged better than the medium. Reposting someone else's accidentally funny remark to an audience primed to appreciate it is now roughly half of what the internet does all day; alt.humor.best-of-usenet did it with a submission queue, a named selector, and the original author's name attached. Of everything about the group, that last habit may be the part that dates it most.