The sci.* hierarchy

How the hierarchy came to be

sci.* arrived with the rest of the Big Seven when the sprawling net.* namespace was reorganised across 1986 and 1987. That reorganisation is told in full on this site’s soc.* page, since the fights it provoked were mostly about groups that ended up elsewhere. What sci.* took from it was a name that made a promise: the old net.physics, net.math and their neighbours moved in under a single label, and the label shaped who posted there for years afterwards.

Creation in the Big Seven was deliberately slow. A proposal went out as a Request For Discussion in news.announce.newgroups, was argued over in news.groups for a set period, and then went to a public Call For Votes. The guidelines that settled into place asked for at least twice as many yes votes as no, and at least a hundred more, tallied by neutral volunteers who published the count. A field had to persuade a general Usenet electorate that it deserved its own room, so the resulting branch map reads less like a faculty list than like a record of which specialisms had enough people online, early enough, to muster a hundred votes.

The second thing that shaped sci.* was the return address. A large share of posters were researchers, postdoctoral fellows and graduate students writing from university and laboratory accounts, under names their colleagues would recognise. That did not make the hierarchy polite, but it made it accountable: an assertion drew a request for the reference, a reference drew someone who had read it, and hand-waving cost more than it did where everyone was a pseudonym.

What lived here

The trunk groups covered the expected ground — sci.math, sci.physics, sci.chem, sci.bio, sci.astro and sci.med — and most grew children as their traffic outgrew a single room: sci.math.num-analysis and sci.math.symbolic under mathematics, sci.med.nutrition and sci.med.physics under medicine, sci.astro.hubble under astronomy. The earth sciences gathered under sci.geo, with geology, meteorology, fluid dynamics and oceanography as separate rooms.

Around the trunk sat the applied and engineering-adjacent groups: sci.engr with its civil, mechanical, chemical, control and biomedical branches, plus sci.electronics, sci.optics, sci.materials, sci.energy and sci.environment. The human sciences held sci.psychology, sci.lang, sci.anthropology, sci.archaeology, sci.logic and sci.econ. sci.agriculture and sci.aquaria carried the practical end, and sci.skeptic sat on the outer boundary, where claims no science group would host were examined, often by people with scientific training.

The research groups and the noise they answered

Open access had a predictable cost. An unmoderated group named for a discipline is an open letterbox, and the ones named for physics and mathematics attracted a steady arrival of unsolicited theory: perpetual motion, relativity overturned, angles trisected, primes explained. The problem was old enough to have its own Usenet vocabulary, and sci.* met it with a mixture of patience and exasperation: long courteous replies from working researchers, then shorter ones, then the crackpot index, a scoring scale proposed in 1992 by the mathematical physicist John Baez and revised in 1998, which circulated widely as shorthand for the genre.

The structural answer was a moderated companion group. sci.math.research and sci.physics.research were both moderated forums for discussion at research level, with articles approved before they appeared, and other fields built the same thing: sci.astro.research, sci.crypt.research. The parent groups stayed open, and stayed noisy, which was the intended division of labour rather than a failure of it.

Two exceptions, and the drift away

Two branches reached well beyond the academy. sci.crypt was one of the places where the public argument over civilian cryptography ran through the 1990s, alongside the mailing lists: the key-escrow proposals, the United States export regime, and the question of how far a cipher could be trusted once examined openly. It was also where the description of the RC4 cipher, leaked anonymously to a cryptography mailing list in September 1994, was reposted and picked apart within days. The wider dispute is documented under the Crypto Wars.

sci.space did the opposite: it brought a lay readership in. Missions were followed there by readers with no professional stake, and the branch subdivided into sci.space.news, sci.space.shuttle, sci.space.policy, sci.space.tech and sci.space.history — a split between reporting, engineering, politics and memory that few other sciences needed.

The decline was a migration rather than a collapse. arXiv, begun in 1991 at Los Alamos to circulate physics preprints, took the function of seeing work early; departmental and society mailing lists took the shop talk; journals put their own contents on the web. By the time general Usenet was folding into spam and binaries, the professional traffic had somewhere better to be. What the archive preserves is narrower and harder to reproduce: a record of how scientists argued in public, with their names attached, in front of anyone who cared to read.