soc.culture.ethiopia.misc

Long-form reference · 6,515 words · about a 28-minute read

A group named after the remainder

Most Usenet groups are named for what they are about. This one is named, in its final component, for what it is not about: .misc, the suffix the Big Eight hierarchies used by custom for the traffic a family of groups could not place anywhere else. soc.culture.ethiopia.misc was the unmoderated half of a pair of Ethiopian groups chartered in 1996, and the only name in the three-name Ethiopian family still carried in the master list of Usenet newsgroups today.

That makes it an unusual page to write, because the group’s own documentary record is short and belongs, almost in its entirety, to its sibling. The 1996 proposals, the drafts, the moderation policy, the ballot and its roll, the forgeries around the plain name and the removal of the moderated room in 2006 are set out document by document on the page for soc.culture.ethiopia.moderated, and are not retold here. What is left for this page is the thing the sibling page cannot cover, because it is not about that group: the suffix itself.

The .misc convention has a real history, an arithmetic that can be counted from the current namespace, a doctrine written down by the people who ran the group-creation process, and at least one archived attempt to turn the custom into a rule. It also has consequences, one of which is visible in this family: of the three Ethiopian names, the catch-all is the survivor. This article establishes what a .misc group was across Usenet, what this particular catch-all’s charter promised in advance, and what the record does and does not show about the room that resulted. Where it can quote a primary document it quotes it; where the record is empty it says so rather than filling the space.

What a name component was supposed to do

Before the suffix, the doctrine. Usenet had two quite separate sets of rules about newsgroup names, and only one of them was enforceable.

The enforceable set is mechanical and short. As stated in the group-creation policies distributed with the master configuration files at the Internet Systems Consortium — criteria based, the document says, on the rules in RFC 5536 — a name component may contain only the characters a-z0-9+_-, must contain at least one non-digit, must not contain uppercase letters, and must begin with a letter or digit; a name must have at least two components; the first component must begin with a letter and must not be control, to or example; the sequences all and ctl must not be used as components at all; and the whole name must be no longer than eighty characters. Two words are forbidden outright as name components anywhere — all and ctl — and three more are barred from the first position. That is the whole of the vocabulary the namespace refuses, and misc appears nowhere in it.

The second set is taxonomic, and it was never a rule in any sense a news server could check. Its classic statement is a 1999 article by David W. Wright, “Guidelines on Usenet Newsgroup Names”, which the Big-8 Management Board still reproduces, in its own words very slightly edited, in its guidance on naming. Wright’s framing is that group names are “structured, hierarchic, taxonomic but not definitive”, and it is the last of those four adjectives that produces catch-alls:

News group names are inclusive rather than definitive. That is to say, a group name defines an area in which a message may be posted if there is no other group with a better name fit. The name does not define exact limits to the group, eliminating subjects that do not exactly match the definition.

The .misc group follows from that sentence as a matter of logic. If every name in a tree describes an area rather than a boundary, then a well-formed tree needs one area whose description is the complement of all the others — a room defined by exclusion, whose criterion for admitting an article is that no better-fitting name exists. Wright’s guidance says the same thing again from the other end, in an instruction to proposers: “Create general groups before creating very specific ones.”

It is worth noting what the Board’s naming page does not contain. It gives explicit guidance on the .moderated suffix — that suffix “should be used when it adds clarity, generally when the group is an alternative or companion group in a hierarchy” — and it works through five numbered examples of naming problems. It offers no equivalent paragraph on .misc. The most common final component in the Big Eight is the one the written guidance never quite gets around to defining.

Where the convention came from

The suffix has no founding document that this article has been able to locate, and the honest account of its origin is that it accreted. What can be dated is the shape of the namespace it grew in.

Until 1987 Usenet groups lived in three loose trees — net.* for ordinary discussion, mod.* for moderated groups, fa.* for traffic gatewayed from ARPANET — with names that had grown by accretion. The Great Renaming of 1987 sorted them into seven topical hierarchies, and that story belongs to the soc.* hierarchy page and is not retold here. What matters for this page is one detail of the result: of the seven trees the Renaming produced — comp, misc, news, rec, sci, soc and talk — one was itself the catch-all. misc.* existed to hold subjects that fitted none of the other six, and it still holds, at its own root, a group whose description in the master list reads “Various discussions not fitting in any other group.” The Big Eight were therefore built, from the first day of the renamed namespace, on the principle that a complete classification needs a bin for the unclassified. The leaf-level suffix is that principle applied one level down, and then applied again, and again.

By the beginning of 1990 the practice was established enough to be complained about. On 1 March 1990 Richard Welty, writing from an address at GE’s research and development site in Niskayuna, New York, posted a call for discussion to news.announce.newgroups and news.groups under the subject “Call for discussion: .misc group creation procedure”. The Internet Systems Consortium still archives it. It opens:

it seems to me that there is a significant loophole in the newsgroup creation procedures as they are currently formulated; they permit niche newsgroups to be created rather easily even when there is no .misc newsgroup in sight.

Welty gave rec.radio.* and rec.models.* as current examples and rec.arts.* as a past one, and proposed a change to the procedure “so that when a new spot in the hierarchy is created, that an appropriate .misc newsgroup is automatically put in place”. His reason for wanting it automatic is the sentence in the posting most useful for understanding why so many trees lacked a catch-all: “while we can run individual votes on specific misc newsgroups (such as has been done in the past), they’re generally not sexy enough or specific enough to gather a lot of votes.” He ended by asking the network whether the idea was reasonable.

That posting establishes three things at once, all of them dated. The convention existed in 1990 and had a recognised name. It was not part of the written procedure. And its principal practical difficulty was already understood: a group defined by what it is not attracts fewer enthusiasts than a group defined by what it is, and the creation rules of the period, as the proposals of the era set them out, required a two-thirds majority and at least a hundred more yes votes than no votes before anything could be created at all.

Welty was not speculating about that difficulty. Six weeks before he posted, on 20 January 1990, the votetaker J.T. Conklin had announced the result of a ballot on exactly such a group: “The newsgroup news.software.misc failed due to lack of interest: 62 Yes votes to 12 No votes.” Fifty more yes votes than no votes, and a five-to-one majority of those who bothered, was not enough. The archive does not record whether that failure prompted Welty’s posting; the two documents simply sit six weeks apart in the same archive, and news.software.misc would not exist for nearly seven more years.

The catch-all that founded a hierarchy

The clearest demonstration that Usenet treated the catch-all as a structural component rather than an afterthought is the creation of the eighth Big-Eight hierarchy, which consisted of creating one .misc group.

A Request for Discussion for humanities.misc was posted on 4 January 1995, over a discussion document signed by Todd Michel McComb and dated 14 December 1994, revised 3 January 1995. The proponent of record, as the vote result names it, was a mailing list: the Arts & Humanities list at [email protected], formed for the purpose. The proposal’s own explanation of why the new hierarchy should be inaugurated by a catch-all is unusually explicit:

We propose that the first step in establishing this hierarchy be the creation of humanities.misc, a second-level “misc” group, as is prevalent throughout the Usenet namespace.

And, in the Call for Votes: “the specific newsgroup humanities.misc serves primarily to inaugurate the humanities.* hierarchy, and is designed in keeping with the existence of *.misc groups in most of the other hierarchies.” The reasoning offered for putting only one group on the ballot was procedural — it let voters decide the hierarchy rather than the merits of any particular topical group, and it avoided the problem of dependent votes — with the topical groups to follow “on individual initiative” once the hierarchy existed. The charter was written in the standard catch-all form: the group “serves explicitly as a forum for those disciplines which do not have dedicated newsgroups”.

One line of that proposal deserves to be quoted on this page in particular, because eighteen months later an Ethiopian proposal would build its whole two-room design on the same premise without citing it. Explaining why humanities.misc would be unmoderated, the document says simply: “This group is proposed as unmoderated due to the futility of moderating a ‘misc’ group.” A room whose admission criterion is that nothing else fits cannot easily be policed for fit.

The ballot closed at 23:59:59 GMT on 5 April 1995 and the result was posted on 11 April: humanities.misc passed 508 votes to 97, with three abstentions. The Big-8 Management Board’s own naming guidance still records the outcome in a single sentence — that in 1995 the creation of humanities.misc, a single newsgroup, also created a new hierarchy and changed the Big-7 into the Big-8. There is a small joke buried in the distribution list, too: the Call for Votes was cross-posted to misc.misc, rec.misc, soc.misc and rec.arts.misc, among others. When Usenet wanted to reach everybody at once, it posted to the catch-alls.

Counting the convention

The convention can be measured, because the Internet Systems Consortium distributes the master newsgroups file — the canonical description of the namespace, updated hourly — and it can simply be counted. The copy read for this article was the one distributed on 25 August 2026. It lists 45,003 groups across every hierarchy the archive tracks, of which 1,898 belong to the Big Eight.

Of those 1,898 Big Eight groups, 191 have misc as their final component. That is a little over one group in ten, and it makes misc by a very wide margin the most common last component in the Big Eight. The next most common is marketplace, with 40; no other final component reaches twenty, and the nearest challengers are tech and announce with 16 each and programmer with 14. Across the whole file, all hierarchies included, 707 group names end in .misc.

The distribution across the eight hierarchies is uneven in the way one would expect, since it tracks how finely each tree was subdivided rather than how fond of catch-alls anybody was:

• comp.* — 90 groups ending in .misc, out of 677
• rec.* — 63, out of 659
• misc.* — 14, out of 100
• sci.* — 7, out of 185
• soc.* — 7, out of 213
• news.* — 5, out of 26
• talk.* — 3, out of 31
• humanities.* — 2, out of 7

Seven of the eight hierarchies carry a catch-all at their own root: comp.misc, misc.misc, news.misc, rec.misc, sci.misc, soc.misc and humanities.misc. The exception is talk.*, which has no talk.misc but does have talk.philosophy.misc, talk.politics.misc and talk.religion.misc — three catch-alls one level down, in a hierarchy of thirty-one groups whose entire purpose was open-ended argument.

How consistently was the convention applied? Counting every prefix in the Big Eight with at least two immediate children — that is, every sub-tree that branches rather than running in a single line — there are 294 such families, and 167 of them, about 57 per cent, have a .misc among those children. Welty’s complaint of 1990 therefore holds up rather well as a description of the namespace thirty-six years later: the catch-all was a strong convention and never a universal one, and roughly two families in five have no bin for the unbinnable.

The convention also nests, to a current maximum of six components. One example is misc.forsale.computers.mac-specific.cards.misc, whose first and last components are both misc: a catch-all for expansion cards, within a catch-all for Macintosh equipment, within the hierarchy that was itself the Renaming’s catch-all. No single document designed that shape; it is what a taxonomy looks like after repeated polite extension by people working one level at a time.

What the descriptions actually say

Each group in the master file carries a one-line description, and the descriptions of the .misc groups are a small genre of their own. Read together they are the best available evidence of what the convention was understood to mean, because they were written independently, by different proponents, across two decades, and they nearly all say the same thing in different words.

Some define the room by exclusion from a named list:

• misc.invest.misc — “Investments not in other misc.invest.* groups.”
• misc.transport.rail.misc — “Rail issues not in other geographic groups.”
• comp.unix.misc — “Various topics that don’t fit other groups.”
• soc.misc — “Socially-oriented topics not in other groups.”
• rec.toys.misc — “Discussion of toys that lack a specific newsgroup.”

Some are addressed directly to the reader holding an article and wondering where to put it. rec.audio.misc, in what may be the most candid newsgroup description in the Big Eight, reads: “Post about audio here if you can’t post anywhere else.”

Some quietly admit that the catch-all is also where the short-lived material goes. sci.misc is described not as general science but as “Short-lived discussions on subjects in the sciences” — a room for the conversation that will not last long enough to justify a group of its own, which is a different job again.

And a few are catch-alls in name whose descriptions read like ordinary topical groups. comp.robotics.misc is simply “All aspects of robots and their applications”, and it has no siblings left in the current list: there is no other comp.robotics.* group for it to be the residue of. That is a state this page will return to.

The remedy for crossposting

Beyond taxonomy, catch-alls were argued for on a practical ground that shows up repeatedly in the group-creation archive: without one, general traffic goes everywhere at once.

The clearest statement of it is a Request for Discussion posted on 1 November 1995 by Jonathan Grobe, proposing a reorganisation of the news.software.* tree — renaming news.software.b to news.software.c-news and news.software.nntp to news.software.inn, and adding news.software.future and news.software.misc. Among the problems the proposal listed with the existing arrangement was this one, stated flatly: “General topics are usually cross-posted because of the lack of a general (or misc) group.” The proposed charter was the standard form once more — “all aspects of news software not better dealt with by more specialized groups”.

The mechanism behind the complaint is worth spelling out, because it is why news administrators cared about catch-alls and not only taxonomists. A Usenet article carried a list of newsgroups in its header, and a poster with a general question about a subject whose tree offered only specialised rooms had one obvious way to reach the right readers: name three or four of the specialised groups at once. Every server then carried the article in all of them, every reader of all of them saw it, and every reply propagated the same way. A single general room converted that fan-out into one copy in one place. The catch-all was, among other things, a bandwidth measure.

The sequel is instructive about how hard these groups were to create even when everyone agreed they were needed. news.software.misc had failed in January 1990. It was proposed again in the 1995 reorganisation, and this time it passed 178 votes to 18, as reported in news.announce.newgroups on 4 December 1996; the control message creating it went out on 10 December 1996, nearly seven years after the first attempt. It is still in the master list today, described as “Usenet-related software other than readers and servers.”

Catch-alls in the culture branch

The soc.culture.* branch used the convention sparingly, and understanding how is necessary before looking at the Ethiopian case, because the branch’s own catch-all does a different job from the one the suffix usually does.

soc.culture.misc exists, and its description in the master list is “Group for discussion about other cultures.” That is not a bin for topics without a room. It is a bin for peoples without a room — the address for a culture that never had the proposer, or the voters, to get a group of its own. The branch’s general logic and its shelf of national and regional rooms are the subject of the soc.culture.* branch page and are not restated here; the relevant point is narrow. The branch’s catch-all was horizontal, catching whole communities that fell between the named groups, rather than the overflow of any particular community.

That distinction matters because it explains why a country with a group of its own might still want a catch-all beneath it. The branch’s design gave each community one room. A community that decided one room was not enough had to build the second level itself, and the second level is where the .misc convention lives.

Very few communities did. The current master list carries 139 groups under soc.culture, of which 17 have four or more components — six regional subdivisions under soc.culture.indian, three subject groups under soc.culture.pakistan, two moderated companions, and six others. Exactly one of the 17 ends in .misc: soc.culture.ethiopia.misc. In the whole of the largest branch of soc.*, it is the only catch-all beneath a country’s name that the list records.

Elsewhere in soc.* the suffix appears in the ordinary way. soc.genealogy.misc is the clearest example, and it arrived by a route that says something about how catch-alls come to exist: when the family-history community subdivided its single old group in 1994, the general room was not abolished but renamed into the catch-all slot, a story told on this directory’s soc.* hierarchy page. A flagship becoming a residue is one of the two normal ways a .misc group is born. The other is the one this page is about — being chartered from the start as the room for what the room next door will not take.

Chartered as a companion, in the same document

soc.culture.ethiopia.misc was proposed in the third and final Request for Discussion of its family, posted on 4 June 1996 under the subject line “3rd RFD: soc.culture.ethiopia.{moderated,misc}”. The two groups shared a document, a proponent, a rationale and, in due course, a ballot paper. They passed together on 8 September 1996, the catch-all by 313 votes to 126; the sequence of drafts that produced the pair, the argument about hierarchy placement that produced the name, the moderation policy attached to the other half and everything else in that file belong to the page for soc.culture.ethiopia.moderated.

What belongs here is the newsgroups line the proposal asked for, which the drafts carried unchanged from June to September:

soc.culture.ethiopia.misc Ethiopian culture. (Unmoderated)

Two rooms, one description between them, distinguished by a single parenthesis. The proposal did not present the catch-all as a lesser room or a fallback. It presented it as the other half of a pair.

The charter, in the proposal’s own words

The charter for the unmoderated group runs to four sentences, set out in three short paragraphs. It is reproduced here in full, exactly as it appears in the third Request for Discussion of 4 June 1996, in the two Calls for Votes, in the result of 8 September 1996, and in the control message that created the group:

Soc.culture.ethiopia.misc will serve as an unmoderated forum in which to debate the politics and political economy of Ethiopia. The charter shall be identical to that of soc.culture.ethiopia.moderated, but there shall be no mechanism to ensure that posters observe the charter. Posters are asked to extend their fellow Usenet readers the respect they deserve by avoiding excessive flamewars, namecalling, and crossposting. Discussion of ethnic issues is on-topic in soc.culture.ethiopia.misc, as are rejected postings from soc.culture.ethiopia.moderated and other appropriate groups.

Four sentences, and each of them does a distinct job.

The first states a subject, and states a narrower one than the group’s eventual description would suggest: the politics and political economy of Ethiopia. This is a residue of the group’s drafting history — in the previous draft the same sentence, with the group’s name changed, had been written for a companion group proposed in the talk.* hierarchy — and it survived the move into soc.culture.* unaltered.

The second sentence is the one that makes this a catch-all rather than a second topical group, and it is a genuinely unusual piece of charter drafting. Rather than writing its own scope, the group adopts its sibling’s charter wholesale — the twelve numbered topics and all — and then removes the enforcement: “the charter shall be identical… but there shall be no mechanism to ensure that posters observe the charter.” Two rooms with the same written constitution, one of which has a door.

The third sentence asks for restraint and provides no means of requiring it. The verb is “asked”, and the sentence immediately before it has already stated that no mechanism exists.

The fourth sentence is the undertaking this page turns on. Ethnic issues are declared on-topic; and so, expressly and in advance, are “rejected postings from soc.culture.ethiopia.moderated and other appropriate groups”. Before either room existed, the pair had agreed in writing what would happen to the material one of them turned away.

What the division of labour was meant to achieve

The record does not contain a statement of purpose for the arrangement beyond the charters themselves. What the charters set out, read together, is a design, and the design can be described without guessing at motives.

Moderation on Usenet worked by interposing a person between the poster and the network: submissions went by mail to a moderator, and returned to the network only carrying an Approved header, without which servers would not propagate them. That machinery has its own account on the soc.* hierarchy page, which is where a reader should go for it. Its cost is structural and was well understood at the time: every rejection is an article that has been written and will not appear, and the person who wrote it has nowhere to put it.

A pair of rooms answers that cost directly. The moderated room can apply a written policy strictly, because rejection is not suppression when the rejected article has a chartered destination one component along. The catch-all can be genuinely unmoderated, because it is not the community’s only public face. And the border between them is declared in advance, in the founding documents, rather than negotiated later during whatever argument first tests it.

The Ethiopian moderation policy makes the design legible. Its five rejection categories, reproduced in full on the sibling page, are mostly not about politics: one concerns the subject matter, one commercial advertising and spam, and the remaining three cover repetition and over-quoting, incivility, and — the category most often noticed by later readers — misspellings, ungrammatical prose and lines longer than seventy-five characters. A policy of that kind will reject material that is perfectly on-topic and simply badly typed. The companion group is where that material was meant to land.

The design also had a precedent the proposers do not cite but which had been stated plainly on Usenet eighteen months earlier: the humanities.* proponents’ observation that moderating a catch-all is futile. If a room’s admission criterion is that nothing else fits, there is very little for a moderator to enforce. Pairing a moderated flagship with an unmoderated catch-all puts the moderator in the room where a moderator has something to do, and leaves the other room alone.

Whether any of this worked is a question the record cannot answer, and this page does not pretend otherwise. Nothing in the material consulted shows how many articles the moderated room rejected, whether any of them reappeared next door, or whether the two groups’ readerships overlapped at all. The design is documented. The operation of the design is not.

The line that a control message wrote in 1996

Big Eight groups were not created by the vote. They were created by a control message — a specially headed Usenet article instructing every news server that received it to add the group — sent after the result by the moderator of news.announce.newgroups, and honoured or ignored by each administrator individually. The Internet Systems Consortium archives those messages, and the file for this group is short enough to describe completely.

It contains four articles. The first is a PGP-signed newgroup message dated 16 September 1996 at 14:30:32 GMT, from David C. Lawrence, citing the 313:126 result, reproducing the charter quoted above, and carrying, beneath the standard heading “For your newsgroups file:”, the description line that gives this section its title:

soc.culture.ethiopia.misc Unmoderated forum to discuss Ethiopian culture.

Two identical messages follow, on 23 September and 16 October 1996, reissued in the ordinary way for servers that had missed the first. The fourth and last article in the file is an rmgroup message of 4 November 2001, sent from a private address rather than by the moderator of news.announce.newgroups, whose entire body reads “please remove the bogus newsgroup soc.culture.ethiopia.misc”. It did not take effect, and it was not aimed at this group in particular: the control archive for news.software.misc records an rmgroup from the same sender fifteen minutes earlier that same night, and three more from another address over the preceding two nights. The 2001 messages and their context are noted on the sibling page, which holds the family’s administrative history.

The description line is the detail worth dwelling on. The proposal had asked for “Ethiopian culture. (Unmoderated)”. What the creating message actually put into the world’s newsgroups files was “Unmoderated forum to discuss Ethiopian culture.” — a rewritten line, of the kind a description has to be when it must stand on its own in a list of tens of thousands. And that rewritten sentence is still there. The master newsgroups file distributed from the Internet Systems Consortium on 25 August 2026 carries, verbatim:

soc.culture.ethiopia.misc Unmoderated forum to discuss Ethiopian culture.

Almost thirty years, unamended, in a file that is rebuilt every hour.

The one that survived

Of the three names in this family, that line is the only one left in the Big Eight namespace.

The moderated room was removed by the Big-8 Management Board in June 2006, on the application of its own moderator; the plain name at the head of the family, soc.culture.ethiopia, was never validly created at all, having reached the world’s servers by forged control messages rather than by any procedure the Big Eight recognised. Both stories, with their dates and documents, are on the moderated group’s page. The consequence is what concerns this page, and it is worth stating without ornament: the catch-all outlived both the room whose rejects it was chartered to receive and the flagship name it was chartered to sit beneath. The subordinate clause survived the sentence.

That outcome is odd but not unique, and the namespace can be asked how odd. Of the 191 .misc groups in the Big Eight, 24 are now the only group anywhere in their sub-tree — catch-alls with nothing left to catch for. Seventeen of the 24 are in comp.*, and they name platforms and products of the 1980s and 1990s: comp.unix.xenix.misc, comp.sys.ibm.sys3x.misc and comp.os.ms-windows.win95.misc among them. Of the rest, humanities.design.misc is one, and two are in soc.*: soc.org.service-clubs.misc, and this one. It is the only such group in soc.culture.*.

One documented removal exercise excluded names of this shape from consideration, and it can be dated precisely. On 1 February 2011 the Big-8 Management Board published a Request for Discussion titled “The Great Downsizing 2011/1”, proposing to remove 198 unmoderated newsgroups. The selection criteria were listed at the head of the document, and they were repeated unchanged in the second Request for Discussion of 3 March, in the Last Call for Comments of 10 April, and in the result of 24 April 2011 that removed all 198. The documents state that every group on the list satisfies four conditions:

• “no moderated groups”
• “no group names matching *.misc”
• “zero on-topic, non-crossposted threads in the past 18 months”
• “on-topic questions that received no on-topic answer do not count”

Emptiness alone, in other words, was not sufficient grounds for removal in that exercise if the name ended in .misc. The documents give no reason for the exemption, and none is supplied here. The board’s earlier low-traffic removal policy, adopted on 2 October 2007 after a drafting process that reached a fifth Request for Discussion, contains no such clause, so the exemption is a feature of the 2011 exercise as documented and not demonstrably a standing rule. What can be said is that in a mass removal of 198 Big Eight groups, a group of exactly this shape was excluded from consideration by name.

None of which is evidence that anyone was reading. Survival in the master list is an administrative fact about a configuration file. The Consortium’s active file, distributed alongside the newsgroups file, records this group as soc.culture.ethiopia.misc 0000000000 0000000001 y — placeholder high and low water marks and a status flag meaning that posting is permitted and the group is unmoderated. It is not an article count, and nothing in that file or any other consulted here reports one.

What a catch-all carried

The genres of traffic that fell to a catch-all in the culture branch can be described, and they should be described carefully, because describing a genre is not the same as describing this group. Nothing in the material consulted for this article documents what was actually posted in soc.culture.ethiopia.misc — not a thread, not a subject line, not a volume, not a period of activity. What follows is about the shape of the residue in general, and it is offered as that.

The first genre is the one the charter named: material a moderated room declined. Under the Ethiopian policy that meant off-charter argument, but also incivility, repetition, over-quoting and bad typing, since four of the five written grounds for rejection in that policy turn on something other than subject matter.

The second is announcements. Where a Usenet family wanted a dedicated room for notices it said so in the name, and the Big Eight still lists 16 groups ending in .announce. Where it did not, announcements went to the general room, and in a family whose only general room is the catch-all, they went there: meetings, appeals, publications, conferences, the ordinary bulletin traffic of an organised community.

The third is classified advertising, for the same structural reason. The Big Eight carries 40 groups ending in .marketplace and a misc.forsale.* tree of sixteen groups, all of it created because for-sale traffic is unwelcome in a discussion group and unstoppable without an alternative. A national community with no marketplace group and no announcements group had one room below the flagship, and it was the .misc.

The fourth is requests — the interrogative traffic that has no natural home in a topical taxonomy because it is defined by the asker’s need rather than the subject: where to find something, whether a rumour is true, who knows someone, how to reach an institution. The general furniture of the culture branch, the recurring genres its groups shared and the FAQs that managed them, is described on the culture branch’s own page.

The fifth is simply the remainder: everything on-topic for the family that nobody thought to name. That is the whole point of the suffix, and it is also why a catch-all is so difficult to describe from the outside. A room defined as the complement of a list has contents that depend entirely on what was on the list.

There is one further genre that the record permits mentioning only as a possibility, and it is mentioned because omitting it would flatter the arrangement. An unmoderated room chartered to accept what a moderated room rejects, in a family whose charter put contested political questions expressly on-topic, is by construction the room where the contested material goes. The 1996 charter says as much in its own fourth sentence. Whether that is what happened here is not established by anything consulted, and this page does not assert it.

What the record does not show

The gaps on this page are large and specific, and collecting them in one place is more useful than qualifying every sentence above.

No traffic figure of any kind is established. There is no article count, no posts-per-day estimate, no subscriber number, no readership survey and no period-by-period activity profile for this group at any point between its creation in September 1996 and the present. The master active file carries a placeholder watermark rather than a count, and no other source consulted supplies one.

No moderator question arises, and it is worth stating rather than passing over. This group had no moderator, by design and by charter; the board procedures that deal with dormant moderated groups and absent moderators could not, and did not, apply to it. The only human role the charter defines is that of a poster who is asked to be civil.

No thread, subject line, poster or argument from inside the group is documented in anything consulted for this article. The standard box above links to a public web archive of Usenet; this page has not surveyed it, has counted nothing in it, and therefore makes no claim about its size, its date range, its composition or its representativeness. A characterisation of a group’s two decades based on whatever an archive search happens to surface first would be unreliable, and the temptation to supply one is precisely why it is being declined in writing.

No FAQ, charter reposting, group-maintained document or periodic posting from this group has been located. No scholarly or journalistic account of it has been located. It attracted, so far as this article can establish, no external documentation of its own whatsoever.

No date of dormancy is established. The group is in the master list today and was in it in 1996; nothing consulted shows when, or whether, it stopped carrying traffic.

Finally, no answer is established to the question a reader is most likely to bring to this page — whether the two-room arrangement did what its designers intended. The design is documented in full. Its operation is not documented at all.

Scope, sources and limits

This article is about a newsgroup name and the convention that produced it. Its primary sources are the archive of news.announce.newgroups and the archive of Usenet control messages, both as mirrored by the Internet Systems Consortium, together with the master newsgroups and active files distributed from the same place; the copy of the newsgroups file read here was the one distributed on 25 August 2026, and every count in this article — the 45,003 groups, the 1,898 in the Big Eight, the 191 catch-alls, the 294 branching families, the 24 solo survivors, the 139 groups in soc.culture — was computed from that file rather than taken from any secondary source. The charter, the newsgroups lines and the control-message text are quoted from those archives verbatim or not at all. Two further documents are quoted from the Big-8 Management Board’s published naming guidance: its statement on the .moderated suffix, and its very slightly edited reproduction of David W. Wright’s 1999 “Guidelines on Usenet Newsgroup Names”.

The neighbouring pages of this directory own the neighbouring subjects, and this article has deliberately not taken them. The whole 1996 documentary record of the Ethiopian family, the ballot and its roll, the forgeries and the 2006 removal belong to soc.culture.ethiopia.moderated. Ethiopia itself — the country, its chronology, the network access figures, the Ethiopic script on the wire, the calendar and the cultural traffic — belongs to soc.culture.ethiopia. The culture branch belongs to its own hub page, and the Great Renaming, the group-creation procedure and the mechanics of moderation belong to the soc.* hierarchy page. The complete directory of groups preserved here is the all-groups list.

On length and shape: this article is long for a group about which almost nothing group-specific is known, because almost everything in it is about the convention, the namespace and the paperwork rather than about the room. Where the material ran out, as it does repeatedly in the two sections above, the article says so rather than continuing.

On the subject matter. The charter quoted here declares the politics and political economy of Ethiopia, and discussion of ethnic issues, to be on-topic in this group, and those questions are matters of serious and continuing disagreement. This page records what a 1996 proposal document said, in what words, and on what date. It takes no position on any question the group’s charter admitted, offers no verdict on any argument that may have been conducted there, and is not a source on the merits of anything its subjects were arguing about.