alt.newbie

Long-form reference · 8,765 words · about a 38-minute read

Thirty-four characters of welcome

The nearest thing Usenet keeps to a register of itself is a pair of plain-text files published by the Internet Systems Consortium. One, called active, lists the group names a server may carry and whether posting to them is permitted. The other, called newsgroups, pairs each name with a single line of description, so that a newsreader can tell a subscriber what a name means before they subscribe to it. In the copies consulted for this page, which list 45,003 names apiece, alt.newbie appears in both: flagged y for open to posting in the first, and in the second carrying its name, two tab characters, and thirty-four characters of description.

The altnet housewarming committee.

That is a friendly sentence, and in this hierarchy a friendly sentence is a minority event. In the alt.* part of that file a description is not editorial work by the archive’s maintainers: the stated rule is that a group’s description becomes whatever the last control message for it carried, so the lines record the temper of whoever last typed one; a great many are jokes at somebody’s expense, and a fair number are simply abuse with a group name attached. This directory’s page for alt.america is built around the opposite case — a description that reads, in full, “Created by a British cluebag; probably now has USA talk.” Same file, same hierarchy, same one-tab-then-text format; one line insults the founder of the group it is describing, and the other offers to help you move in.

The metaphor is doing real work. A housewarming is what neighbours hold for someone who has just arrived, and a committee is a standing body rather than an occasion: not a welcome, but an arrangement for welcoming, indefinitely renewed. The prefix altnet treats the alt.* hierarchy as a network in its own right rather than a branch of somebody else’s — which, in the sense that its groups propagated only where administrators chose to honour them, it very nearly was. The sibling name alt.newbies, created six days after alt.newbie and still carried in the same file, is described in the same register: “Housewarming for a group.”

The line is data, not editorial prose. It travels with control messages, it is read by software, and its length is governed by a house style intended to keep the whole line inside the width of a terminal screen; the mechanics of how those lines are written, and how rarely they are corrected afterwards, belong to the alt.america page and are not repeated here. What matters on this page is narrower. For a person who had just been given a news account and was scrolling a list of tens of thousands of names, that sentence was very often the entire orientation the system offered about this group before they subscribed to it.

The paperwork, and the joke inside it

Big-Eight groups leave a thick paper trail: a request for discussion, a call for votes, a tally with every voter named, a result, and a control message. alt.* groups leave a control message, if that. The Internet Systems Consortium mirrors the newgroup and rmgroup messages seen on the network since the early 1990s, filed by hierarchy and by group name, and for most alt.* groups this is the only birth record in existence. For alt.newbie the file holds four messages. The first is the creation.

It is dated 15 July 1992 at 21:02:54 GMT, was sent from a netcom.com account by a poster giving her name as Rachel Dafni Christy, and carries the self-approval that alt control messages conventionally do. Its Newsgroups header names two destinations: control, which is where a control message belongs, and talk.bizarre, which is not — talk.bizarre being the hierarchy’s standing home for the deliberately absurd. The message sets out a name, a status of unmoderated, and a description reading “Forum for discussion of newbies.” Then it gives a charter.

The charter is a parody. It takes the boilerplate of a serious group proposal — the sober list of the kinds of article the proponent expects — and applies it, deadpan, to a subject that cannot possibly sustain it:

• “Any and all questions concerning newbies.”
• “Announcements of conferences, conventions, etc.”
• “Product descriptions”
• “News of public service offerings”
• “Research findings”

It then adds, in the voice of a proponent whose group has not yet received a single article: “It is fully expected that the range of discussion will quickly result in the need for an expanded hierarchy under alt.newbie. We will cross that bridge when we get to it and use experience gained in this forum to proceed.” The message is signed off with a title — “President, League of Newbies” — and a signature block quoting the poster’s own remark about parties.

Two things follow from this, and they should be kept apart. The first is that the founding document of alt.newbie is a joke: a satire of newgroup paperwork, addressed partly to an audience of connoisseurs of that satire, proposing a forum for the discussion of newbies rather than a forum for them. The second is that the joke does not settle what the group became. In alt.*, a creator’s intention was one input among many; a name propagated, the name meant something obvious to anyone reading it, and what accumulated underneath was decided by whoever posted there afterwards. A group called alt.newbie, listed in every server’s group table as a housewarming committee, is going to receive questions from beginners whatever its charter says. The honest position is that the paperwork records the founding gesture, the description records how the group was presented to readers for the next three decades, and the traffic in between is not preserved in any source consulted for this page.

One consequence bears directly on the line quoted at the top of this page, and it is worth stating exactly. The description in the 1992 message is not the description in the newsgroups file. Under the rule described in the previous section, a control message supplies a description only by including a line reading For your newsgroups file: followed by the text. None of the four archived messages for alt.newbie contains such a line, and none contains the housewarming sentence; the 2004 message that created alt.newbie.dowap does contain one, which is why that group’s entry in the file matches its own paperwork word for word. Where the housewarming line came from, and when, the public record does not show.

Removed, restored, and never quite settled

The remaining three messages in the archive are a compressed lesson in how alt.* governed itself, which is to say how it did not.

On 18 July 1994 a poster whose organisation line gave a consultancy in Vineyard Haven, Massachusetts, issued a newgroup for alt.newbie with the distribution set to a single local network and a one-line body explaining that the point was to update that network’s active file. It is not a claim on the wider network at all; it is one administrator adding a group to one spool, captured by the archive because the archive collects what it sees. Control messages of this kind are a reminder that group creation on Usenet was never a single event with a date. It was thousands of separate decisions, most of them invisible.

On 22 May 1995 an rmgroup arrived from a poster at the University of St Andrews, sent to alt.config and its control counterpart, asking servers to delete the group. Its stated grounds are the standard formula of the alt.* cleanup campaigns of that period, and it is quoted here in full because the formula itself is the interesting part:

This newsgroup, after three weeks discussion on alt.config, has been selected to be removed as a duplicated newsgroup, a misspelled newsgroup with a duplicate existing, or a joke newsgroup with little or no traffic.

The sentence offers three possible reasons and declines to say which applies. It could not usefully say: the wording is boilerplate, written to cover a run of removals rather than one group, and the sender’s signature block ended with the slogan “Support the rmgrouping of all silly, unused, or duplicated alt.* groups!” Whether alt.newbie was thought a duplicate of alt.newbies, a joke group, or merely quiet is not recorded. What the message does establish is that by mid-1995 the group had been discussed for three weeks in alt.config and had not persuaded that discussion it was worth keeping.

It made no difference. Less than two months later, on 13 July 1995, a newgroup for alt.newbie went out from a Buffalo internet provider with a body of exactly one sentence: “Group missing from sprintlink spool but extent at uunet.” The spelling of extant is the original’s. That is the whole of the reasoning — one large feed had the group, another did not, and an administrator sent a message to close the gap. An rmgroup on Usenet was a request, and most administrators configured their servers to ignore rmgroups entirely, which is why the hierarchy accumulated names rather than shedding them; the mechanism is set out on the alt.* hub and is the reason a removed group could be back in general circulation before the summer was out.

alt.newbies had a shorter argument and lost it. It was created on 21 July 1992, six days after alt.newbie, by a message relayed through a research centre at Columbia University whose body was empty. It was removed on 14 July 1995 — the day after alt.newbie’s restoration — by an rmgroup whose entire justification ran: “redundant, use alt.newbie instead”. Both names are nevertheless still in the current active file, thirty-one years later, both flagged as open to posting. Neither the removals nor the restorations resolved anything; they simply left both names in the register, which is the ordinary end state of an alt.* dispute.

One later descendant is on file. In January 2004 a group called alt.newbie.dowap was created, its newsgroups-file line reading “Another 85 to go...” and its charter describing itself as a forum about one named poster “and his ability to maintain clueless ‘newbie’ status, even after several years on the usenet”. That is not a beginners’ group. It is the other sense of the word, which is the subject of the next section.

Where the word came from, and what it was used for

Newbie reached the network already carrying baggage. Its origin is uncertain: the standard accounts trace it either to the nineteenth-century word newie, or to British public-school slang for a new boy, and record a limited currency among United States troops in the 1960s and 1970s for a new arrival in a unit. It was not coined online. It was adopted online, and the adoption is dated to Usenet: the Jargon File, the long-maintained lexicon of hacker slang, glosses the term as “A Usenet neophyte” and states that it surfaced in the newsgroup talk.bizarre — the same group named in alt.newbie’s founding control message, which is a coincidence worth recording and not worth building anything on.

The Jargon File’s entry then does something more useful than an etymology. It sets out how the label was actually deployed:

Criteria for being considered a newbie vary wildly; a person can be called a newbie in one newsgroup while remaining a respected regular in another. The label newbie is sometimes applied as a serious insult to a person who has been around Usenet for a long time but who carefully hides all evidence of having a clue.

Two properties of the word follow. It was relative — a function of a particular group’s idea of competence rather than of elapsed time — and it was assigned by other people. Nobody described themselves as a newbie except in apology. The entry also notes that the collocation “clueless newbie” was especially common, and the group list bears that out: alt.clueless-newbie-of-the-month exists in the current file, with a description reading “and teh winner is...”, misspelling included. So does alt.newbie.dowap, whose charter is the Jargon File’s “serious insult” sense written out as a group proposal.

This matters for reading a group like alt.newbie honestly. Its name is not a neutral category label; it is a word that in most of its uses on Usenet was mildly to severely pejorative, applied to a person by somebody who had decided they did not yet belong. A group that takes that word for its own name and pairs it with a description about housewarming is doing something specific: claiming the label as an address rather than an accusation. Whether the traffic inside lived up to that is not documented. The naming is.

Why Usenet was hard to enter

Usenet was hostile to newcomers by structure before it was hostile by temperament, and the structural part is the part that can be described precisely. The medium itself is covered on this site’s introduction to Usenet; what follows is only the subset of its architecture that bore on a person arriving for the first time.

There was no front door. Usenet had no home page, no registration, no terms of service, no owner and no central address; it was a protocol for passing articles between independently administered servers. Access came from an institution or a provider, along with whatever newsreader that institution had installed and configured. A new user did not choose the software, was rarely told who ran the server, and had no obvious person to ask. The one piece of orientation the system produced by itself was the group table — tens of thousands of names, each with a line of description that the file’s own house style asks to keep inside fifty-five characters, written by a stranger years earlier, sometimes a joke, sometimes an insult, sometimes the words “No description.”

Nothing in the software enforced any convention. A newsreader would let a beginner quote three hundred lines in order to add “me too”, crosspost to forty groups, post a test message into a live discussion, or set a signature block longer than the article, and none of these produced a warning, because there was nobody in a position to issue one. The conventions existed entirely as social expectations held by the people already there, and they were enforced the only way an ungoverned network can enforce anything: by somebody telling you off in public, at length, in front of everyone.

Three technical properties made that badly asymmetric. The first is reach: an article went to every server that carried the group, which for a widely propagated group meant most of the network, and it went there automatically. The second is irreversibility. RFC 1855, the netiquette document discussed below, puts it flatly — “In general, it’s not possible to retrieve messages once you have sent them. Even your system administrator will not be able to get a message back once you have sent it.” A cancel message existed, but like every other control message it was only a request, and honoured unevenly. The third is that the order in which articles arrived was not the order in which they were sent. Because news propagated hop by hop between servers, the same RFC warns that “it is possible to see a response to a message before seeing the original” — which is precisely why quoting mattered, and why a reply that quoted nothing was not merely bad manners but genuinely unreadable at half the sites receiving it.

To this the era added money. Articles moved over dial-up links and leased lines that somebody paid for, often by the minute or by the byte, and a full news feed was a real line item in a university’s or a company’s budget. RFC 1855’s advice to keep a signature to four lines is justified in the document itself on exactly that ground: “Remember that many people pay for connectivity by the minute, and the longer your message is, the more they pay.” A beginner’s wasted bandwidth was other people’s wasted money, and this is the substance behind a great deal of what otherwise reads as pure snobbery. The satire of the period made the arithmetic explicit, calculating the collective hours consumed by tens of thousands of readers each skipping past one unnecessary article.

Finally, the standard remedy was itself a barrier. RFC 1855’s general guidelines for mailing lists and newsgroups open with it, before anything about quoting or crossposting: “Read both mailing lists and newsgroups for one to two months before you post anything. This helps you to get an understanding of the culture of the group.” That is good advice and it is also a demand for two months of silence from somebody who has just been given an account and has a question now. A group whose description promises a housewarming is, among other things, a proposal to shorten that interval to nothing.

September, and then September again

The rhythm that produced the problem is well documented and frequently misdescribed. For most of the 1980s and early 1990s, Usenet’s intake was seasonal. New users arrived slowly and singly for most of the year, and then in September large cohorts of first-year students in North America were given university accounts and reached the network all at once, in numbers that briefly overwhelmed the ordinary process by which a group absorbed a newcomer. The influx was a standing joke and a standing irritation, and it ended: by the winter the intake had been assimilated or had drifted away, and the groups returned to their usual composition. The seasonality is the whole point. Usenet’s method of inducting new arrivals was not a document or an institution. It was a ratio, and the ratio recovered every year.

What changed at the start of the 1990s was the supply. Commercial services began selling network access to the general public and including Usenet in the package: Delphi, a dial-up service running since 1983, provided national consumer Internet access in 1992 with e-mail, FTP, Telnet, Gopher and Usenet among its features, and the far larger America Online followed. These users did not arrive in September. They arrived on the day they signed up, continuously, in numbers that did not fall off, and the ratio never recovered.

The phrase that fixed this in the language was assembled over a few weeks in early 1994, in newsgroups devoted to folklore. On 8 January 1994 a poster crossposted a complaint asking whether Delphi had unleashed a staggering number of oddities onto the net, and drew the reply that it was of course perpetually September for Delphi users. The day before, the same message posted to a second folklore group had drawn a joke about a “Never-Ending-September”. On 25 January, in alt.folklore.computers, the poster who had coined that phrase wrote the sentence usually cited as the origin: “It’s moot now. September 1993 will go down in net history as the September that never ended.” The exact wording Eternal September appears to be slightly later, in a February 1994 posting by a third writer, who used it as the subject line of a complaint about article quality and thereby named the thing.

The standard accounts differ on one detail, and the difference is worth stating rather than smoothing over. The Jargon File’s entry says that “in September 1993, AOL users became able to post to Usenet”. The encyclopaedic account, following Wendy Grossman’s 1997 book Net.wars, dates the opening of AOL’s Usenet gateway to March 1994 and treats the September 1993 date as marking the broader commercial influx of which AOL was the largest single component. Both agree on the mechanism and on the season; they disagree by about six months on which provider crossed which threshold when. A page that tells you confidently which is right is overstating what the record supports.

The consequence for a group like this one is structural rather than atmospheric. Before 1993, induction was a recurring seasonal task that a group’s regulars performed for a few weeks and then stopped. Afterwards it was a permanent condition, present in every group, every week, forever, and no longer soluble by waiting. That is the environment in which standing arrangements for newcomers stop being a joke about paperwork and start being an obvious thing for a network to want — and it is also the environment in which the patience required to run one stopped being seasonal too.

The official machinery, and what it could not do

Usenet did not lack documentation. It had a moderated group, news.announce.newusers, into which the explanatory articles a newcomer was expected to read were posted on a repeating cycle; it had news.newusers.questions for asking; it had news.answers and a family of hierarchy-level companions such as soc.answers into which every periodic informational posting on the network was crossposted; and it had an archive at MIT from which any of those documents could be retrieved by FTP or by e-mail. That apparatus, its history, its moderators and its long decline belong to this directory’s news.* hub, and the FAQ as a genre and the census of the answers groups belong to soc.answers. Neither is retold here. What belongs on this page is the narrower question of what such machinery can and cannot do for one specific person.

A repository answers a question that has been asked in the right words. It is very poor at three things a beginner needs.

• Discovery. Reading news.announce.newusers requires knowing that news.announce.newusers exists. Nothing in the act of getting a news account told you so. The group appeared in the same undifferentiated list of tens of thousands of names as everything else, with the description “Explanatory postings for new users. (Moderated)”, at a position in the alphabet a very long way from the top.
• Vocabulary. Retrieving a document required knowing its vocabulary. The archive was organised by an Archive-name header — a beginner needed to know the phrase, then the scheme, then how to address a mail server, in order to be sent a file about how to behave. The word netiquette is only searchable by somebody who has already heard it.
• Dialogue. A moderated repository cannot be asked anything. Its articles arrive on a schedule and answer the questions their authors anticipated. The commonest predicament of a new user — not knowing which of one’s many confusions is the one to name first — is exactly the predicament a document cannot resolve, because resolving it requires a person to ask what you are trying to do.

There is a bind in the last of those, and it is structural rather than anecdotal. The standard Usenet remedy for a group filling with traffic it does not want is moderation: submissions go to a moderator, who approves or rejects them. Applied to a group that exists in order to receive beginners’ first articles, that remedy lands on the wrong people first, because a rejection notice is then the network’s reply to somebody who has just worked out how to post at all. news.newusers.questions is carried as a moderated group in the current file; the administrative history behind that flag belongs to the news.* hub. The point on this page is narrower — the only tool the medium offered for keeping such a group on topic was the same tool that turns a beginner away.

Which leaves the gap that a group like alt.newbie sits in. Not the absence of documents — there were more documents than anybody could read — but the absence of anyone to ask which document, and in what order, and whether the thing you are worried about is a real problem at all. That question can only be answered by a person, in public, repeatedly, for free.

Netiquette as a written thing

The conventions were unwritten in the places beginners looked, but they were not unwritten. By the middle 1990s there was a small canon, and it is worth naming precisely, because the authorship is checkable and the documents themselves are the best surviving evidence of what beginners were actually getting wrong.

The formal one is RFC 1855, Netiquette Guidelines, published in October 1995. Its author is Sally Hambridge of Intel Corporation; it is also numbered FYI 28, it is categorised Informational rather than as a standard of any kind, and it states that it is the product of the Responsible Use of the Network working group of the IETF. Its abstract describes it as “a minimum set of guidelines for Network Etiquette” which organisations may adapt for their own use, deliberately written in bulleted form to make adaptation and lookup easier. Its introduction is blunt about why it exists at all:

In the past, the population of people using the Internet had “grown up” with the Internet, were technically minded, and understood the nature of the transport and the protocols. Today, the community of Internet users includes people who are new to the environment. These “Newbies” are unfamiliar with the culture and don’t need to know about transport and protocols.

The scare quotes are the document’s own. A standards body putting the word into an RFC in 1995, in the same year that the commercial influx became the ordinary condition of the network, is a fair marker of how completely the problem had escaped the newsgroups and become an industry matter. The document is organised into one-to-one communication, one-to-many communication, and information services; the NetNews material sits in the middle section, and it is the section a beginner needed.

Alongside it stood a commercial book — Virginia Shea’s Netiquette, published by Albion Books in San Francisco in 1994 — and, more importantly for Usenet itself, the periodic postings. RFC 1855’s own bibliography names four of them with their authorship, which is the most compact verified roll-call of the genre available:

• “Answers to Frequently Asked Questions about Usenet”, original author Jerry Schwarz.
• “A Primer on How to Work With the Usenet Community”, original author Chuq Von Rospach.
• “Rules for posting to Usenet”, original author Gene Spafford.
• “Emily Postnews Answers Your Questions on Netiquette”, original author Brad Templeton.

All four are listed as maintained by the same person, Mark Moraes, at a single address, and each is cited by its Archive-name rather than by any location a reader could visit. That is a fair picture of the whole institution in one bibliography entry: four foundational documents, four original authors who had by then moved on, one volunteer keeping the set current, and a retrieval scheme legible only to people who already knew it. The netiquette that a newcomer was expected to have absorbed was, in practice, a small pile of text files with a single maintainer.

Dear Emily Postnews, and teaching by inversion

The fourth item on that list is the odd one, and it is the one people remember. “Emily Postnews Answers Your Questions on Netiquette” is a satire in the form of an agony column. Emily Postnews, introduced as the foremost authority on proper net behaviour, receives questions from correspondents with names like verbose@noisy and forgetful@myvax, and answers every one of them with the worst possible advice, delivered with complete composure.

The pun is explained in the document’s own closing footnote, which is worth having because both halves have aged out of common knowledge: Emily Post was in her day one of the foremost authorities on etiquette, and postnews was the name of the program used to post an article in earlier versions of the Usenet software. The joke is a compound of an American manners columnist and a command-line utility.

What it mocks is a precise inventory of the things beginners did, each stated as an instruction. On signatures: make yours as long as you can, include a large picture drawn in ASCII characters, and give your telephone number and street address in every article, since people are always responding to Usenet articles with calls and letters. On correcting an error: post at once, use the most confrontational language available, and certainly do not wait a day to see whether somebody else has already done it. On choosing groups: pick as many as you can, expand the list on every follow-up, and never set a Followup-To header, in case somebody misses part of the valuable discussion in the fringe groups. On test messages: test the entire net rather than a subnet, and add the words “please ignore”, since everybody always skips a message with a line like that. On quoting: include the entire text of the article you are rebutting, signature and all, and pack your comments between the lines. On rebuttal generally: compare the poster to Hitler, which the column recommends as “a novel, underutilized rhetorical technique”.

The extended example is the one usually quoted. Asked how many groups an article about an ice-hockey trade should go to, the column reasons its way from the obvious sports group to a news administration group (it is news), to a group about fluid geology (the team is called the Oilers), to an astronomy group and a space group (the player is a star), to a telecommunications group (he was born in the birthplace of the telephone), to a non-existent regional culture group, to news.groups to propose creating it, and finally to talk.bizarre on the grounds that the resulting article will be quite bizarre. It is an accurate parody of a real habit, and it is funny in the specific way that only somebody who has read a great deal of Usenet can be.

It is also, considered as a teaching document, a strange instrument. Satire by inversion works only on a reader who can already detect the inversion. A person who genuinely did not know that crossposting to nine groups was rude could read the ice-hockey passage as an overenthusiastic instruction rather than as a joke about one. The same limitation applies to the other great parody of the era, B1FF — the Jargon File’s “most famous pseudo, and the prototypical newbie”, an invented persona posting in all capitals with cute misspellings, doubled signature blocks and unbounded naivety, originally created by Joe Talmadge, whose BIFF filter was afterwards passed to Richard Sexton, who posted its output more widely. B1FF made the newcomer’s mistakes legible by exaggerating them into a character. Whether it made them correctable by the people making them is a different question, and one the humour is not designed to answer.

Which is the argument for a group rather than a document. A satire and a manual both address the reader as somebody who can already tell which end of the joke they are on. A person answering a question in public addresses the person actually asking it.

The conventions, as conventions of that medium

The specific things beginners broke were not arbitrary rules of politeness. Each of them followed from a property of how news was stored, moved and read, and each of them stopped making obvious sense once that property went away. Four are worth setting out.

Quoting

The convention was to include the part of the previous article you were responding to, marked with a greater-than sign at the start of each quoted line, preceded by an attribution line naming its author, and trimmed to the minimum that made your reply intelligible. RFC 1855 states both halves: include “just enough text of the original to give a context”, because a reader may see the response before the original; and “do not include the entire original”, since “it is extremely bad form to simply reply to a message by including all the previous message”. Its NetNews section adds the rule of thumb that the content of a follow-up should exceed the quoted content, and asks posters to avoid “me too” articles altogether. The nesting of greater-than signs did real work: the number of markers at the start of a line told you how many rounds back that sentence came from, in a medium where nothing else did.

The failure mode was the article that quoted four hundred lines to add one, which every reader on the network then paid to receive, store and scroll past. This is the convention that dates most obviously: it presumes plain text, a fixed-width display, no threading in the reader worth relying on, and a real cost per byte.

Crossposting

A Usenet article could name several groups in a single Newsgroups header. Doing so sent one copy, which each server stored once and showed to a reader once regardless of how many of the named groups they subscribed to. Posting the same text separately to each group — multiposting — sent several copies, which were stored several times and shown to a subscriber of three of those groups three times. The netiquette is therefore counterintuitive and RFC 1855 states it in capitals: if an article really is of interest to more than one group, “be sure to CROSSPOST the article rather than individually post it to those groups”. The same passage supplies the limit — “in general, probably only five-to-six groups will have similar enough interests to warrant this” — and asks that a poster who introduces crossposting to a group warn its readers, since they will otherwise assume that follow-ups stay where they were.

The sin, then, was not crossposting as such but crossposting widely and without direction: naming a dozen groups with no Followup-To, so that a discussion begun in one place proceeded simultaneously in twelve, in twelve incompatible tempers, with no participant seeing the same conversation as any other. Emily Postnews’s ice-hockey routine is a parody of exactly this, and the reason it lands is that the underlying mechanism rewarded restraint in a way no interface ever explained.

Test posts

A new user’s first genuine question about the medium is whether their posting works at all, and the answer is not visible from their own screen. The convention was to find out in a group designated for the purpose. The current group list contains 541 names with a component called test. The two canonical ones carry descriptions written to be unmistakable: misc.test is “For testing of network software. Very boring.” and alt.test is “For testing news software, servers, etc.”

The sprawl of the rest documents the friction better than any account of it could. alt.test.test is described “Think, people! How many test groups do we really need?”; alt.test.two, “This is getting ridiculous.”; alt.test.only, “Stop interrupting testing with discussion.”; alt.test.yer.posts, “Directions for people who find ‘alt.test’ too confusing.”; alt.test.hello-world, more gently, “Human-to-human initial contacts on the net”. One group whose name is itself an insult carries the description “For complete idiots to be told to test in”. These lines were written by irritated people over a decade, and they are the closest thing in the canonical file to a transcript of the reception a misplaced test message received.

Top-posting

The last of the four is the one that outlived Usenet and became a permanent low-grade quarrel in e-mail. In the posting style that the discussion groups had settled on, a reply was interleaved: the original was broken into pieces and each piece answered underneath, or, for a single point, appended below the quoted passage. Top-posting places the reply above an untrimmed copy of the whole original — sometimes abbreviated TOFU, for text over, fullquote under, and sometimes called the Jeopardy! reply style, after the quiz show whose clue-and-response format likewise puts the answers before the questions. The standard mockery of it is a four-line exchange printed in reverse order, whose first line reads “Because it messes up the order in which people normally read text” and whose second asks why top-posting is such a bad thing.

The argument was and is genuinely two-sided, and this page records it rather than settling it. Interleaving keeps each answer next to what it answers and forces the quoted material to be trimmed; top-posting preserves an apparent transcript and puts the newest text where a reader looking for it will find it first. The distribution of the two was never uniform on Usenet either: the comp.lang.* groups, comp.lang.c and comp.lang.c++ in particular, were still insisting on interleaved replies as of the 2010s, while the alt.* hierarchy — this group’s own — was comparatively tolerant of top-posting. What settled the question outside Usenet was software: mail programs that placed the cursor above a copy of the original by default, and later mobile clients that download the beginning of a message first, made top-posting the path of least resistance for the overwhelming majority of people who had never heard of the argument.

Two smaller conventions belong in the same frame, since they turn up in every list of beginners’ errors. Writing in capitals — RFC 1855 puts it as “Use mixed case. UPPER CASE LOOKS AS IF YOU’RE SHOUTING” — and the four-line signature limit, justified in the document by the cost of connectivity rather than by taste. Both are conventions of a plain-text medium with no formatting and a metered link, and both were still being explained to newcomers long after the reasons had stopped applying.

Who staffs a group like this, and why that is unusual

Every voluntary community depends on unpaid effort, but a group whose subject is the questions of newcomers depends on it in a particular and unforgiving way, and the structure is worth stating plainly because it is the group’s real distinguishing feature.

In an ordinary discussion group, the effort a regular puts in accumulates. Threads build on earlier threads, participants come to know each other, and expertise offered in one exchange is available in the next; the reward for staying is a conversation that gets better. A beginners’ group inverts every part of that. Its subject matter is fixed and it does not develop. Its population turns over continuously by design — success is a member who stops needing the group. Its questions arrive in the same small set, indefinitely, and the person answering them derives no intellectual return from the hundredth repetition that they did not already have after the third. What is being supplied is not knowledge. Nearly all the knowledge required was written down years earlier and is a few lines long. What is being supplied is patience, in public, to a stranger, without visible reward, on a schedule set entirely by other people’s arrival.

The obvious engineering answer to a repeated question is to write the answer down once, and Usenet took that answer seriously: it is the entire premise of the periodic posting and of the answers groups. But writing it down relocates the labour rather than removing it. Somebody must still notice that a question has been asked, judge which document applies, point at it in a way that does not read as a rebuke, and — the part no document performs — answer the second question, which is usually the real one. Where the pointing is done briskly it reads as dismissal, and “read the FAQ” became a stock unfriendly reply. RFC 1855 records the harsher version of the same reflex, warning that asking a group about something available elsewhere “generates grumpy ‘RTFM’ messages” and glossing the acronym for readers who had not met it.

Burnout is therefore the standing failure mode rather than an accident, and the Usenet record contains the shape of it even where individual cases are not documented. The machinery of induction was held together by very small numbers of volunteers — the four foundational netiquette documents listed in RFC 1855 in 1995 all being maintained by one person is a fair illustration — and arrangements of that shape end quietly. An institution that depends on one person’s continuing willingness does not usually announce its ending. It just stops posting, and it takes a while for anyone to notice, because the people best placed to notice are the ones who no longer need it. What became of the official new-user group and its repeating postings belongs to the news.* hub rather than to this page.

Nothing about the specific volunteers of alt.newbie is recorded in any source consulted for this page, and nothing is invented here. The structural point stands without them: a group of this kind cannot be sustained by the mechanisms that sustain ordinary groups, because the thing it produces is consumed entirely by people who then leave.

The newbie namespace

One further piece of evidence that the problem was structural rather than local is that almost every part of the network built its own answer to it, independently, in its own language. The current group list is a census of those attempts. Twenty-one names contain the string newbie and twenty-nine contain newuser or newusers, spread across the Big Eight, the alt.* hierarchy, national hierarchies, corporate hierarchies and gatewayed mailing lists.

The official pair sits in news.*: news.announce.newusers, “Explanatory postings for new users. (Moderated)”, and news.newusers.questions, “Q & A for new users of Usenet. (Moderated)”. Around them, the same idea recurs with local wording:

• alt.usenet.newbies — “A place for newbies to ask questions”, which is the plainest statement of the job anywhere in the file.
• alt.comp.newusers — “New user information for the alt.comp.* sub-hierarchy”, a branch building its own front desk.
• demon.ip.support.newuser — “For the hesitant new user to make postings.”, a British provider’s hierarchy conceding, in one adjective, that hesitation was the actual obstacle.
• wa.newusers — “New USENET users come here for assistance.”, and england.news.newusers — “Help and support for newusers.”: the same service offered regionally.
• de.newusers.questions and de.newusers.infos in the German hierarchy, both moderated — respectively “Hilfe zum Usenet” and “Infos und periodische Postings fuer neue User”, the second of which is the periodic-posting institution reproduced wholesale in another language.
• nihon.beginners.salon — “Netnews beginners’ salon.”; nctu.newusers.questions — “Questions from Newcomers”; medlux.newusers — “Q&A of new users”; relcom.newusers — “Q&A of new Relcom users”; bih.news.newusers — “Informacije novim korisnicima”; chinese.newsgroups.newusers — “Postings by new users.”
• vmsnet.announce.newusers — “Orientation info for new users. (Moderated)”, for the VMS community; comp.sys.amiga.introduction — “Group for newcomers to Amigas.”; alt.sys.mac.newuser-help — “A forum for Macintosh users new to the internet.”
• And, in the gatewayed mailing lists, the same instinct in software communities: muc.lists.freebsd.newbies, “Gathering place for new users”; pgsql.novice, “Safe Ground for New Users of PostgreSQL”.

Two entries in that list carry their own commentary. One is free.fr.rec.chasse.newbies, whose description — “About newbies elimination in free.fr.*” — is a hunting pun and a reminder that the welcoming impulse was never the only one in the namespace. The other is eternal-september.newusers, described simply “New to e-s.”: a group for beginners on a free public news server that took its name from the phenomenon described above. A server named for the moment the induction problem became permanent, with its own group for the people arriving.

Against that background, alt.newbie is not a curiosity. It is the alt.* entry in a list that every hierarchy on the network eventually wrote, with the difference that alt.* required no vote to write it and got there in July 1992, and that the sentence it ended up with in the group table was kinder than most.

What replaced it

The function did not disappear; the venue and the economics did. Three changes, in order, and the last two postdate the gateway whose snapshot this directory preserves.

The first was the web forum. Early web-based discussion software dates to 1994, with the W3 Consortium’s World-Wide Web Interactive Talk among the first, and by the turn of the century packaged forum software was widely and often freely available: the free package phpBB was begun by James Atkinson in June 2000 and released as version 1.0.0 that December. Structurally this changed one thing above all others. A forum has an owner. Somebody holds the database, appoints moderators, writes rules that are displayed on the site, edits and deletes posts, and can put a permanent thread at the top of a board saying what the board is for and where beginners should start. None of those actions had any equivalent on Usenet, where no one held the articles, moderation was a matter of an approval header on individual submissions, and the group’s own description was a line in a file that the group could not edit. Over the first half of the 2000s a great deal of hobbyist discussion moved onto software of that kind, where a pinned thread and a named moderator did explicitly and permanently what a Usenet group could attempt only by custom.

The second was onboarding as a designed product feature. Stack Overflow, opened to a private beta in July 2008 and to the public that September, is the clearest case because it addressed exactly the material a beginners’ group handled and did so in software: a reputation score that unlocks abilities incrementally, badges awarded for particular acts, a mechanism for closing a question as a duplicate of an existing one, and an explicit apparatus for editing other people’s questions into answerable form. Where Usenet answered a repeated question by finding a volunteer with the patience to answer it again, a system of this kind answers it by routing the asker to the previous answer automatically, and pays the people who tend the routing in status. The mechanism has been contested on its own site throughout — the duplicate-and-closing apparatus was overhauled in 2013 after sustained argument about it — but the shift is real: induction stopped being an act of individual goodwill and became a feature with owners, metrics and a budget.

The third was scale and instrumentation. Platforms of the following decade can measure what a newcomer does in their first hour, test alternative first-run experiences against each other, and employ staff whose job is the retention of new participants. Usenet could not measure anything. It had no accounts, no central server, no way to know who was reading, and the readership estimates that survive from the period were produced by voluntary sampling extrapolated across the network. A network with no way to count its arrivals has no way to discover that it is losing them, which is one reason the induction problem was addressed for twenty years by hand and by exhortation.

What became of the record is a separate matter. Usenet articles are still archived and the surviving groups are still carried by public news servers; Google Groups, which had long provided a web gateway for both reading and posting to Usenet, retired that function in February 2024. Access to what remains is covered elsewhere on this site. The point for this page is that the reading of an old beginners’ group has itself become a specialist act, requiring exactly the kind of orientation the group was created to provide.

What the record does not show

This page has been assembled from documents about the group rather than from the group. The distinction matters and the gaps are worth listing.

• No traffic of any kind is quoted here. No thread from alt.newbie has been read for this page, no posting is quoted, and no participant is named. Where an archive of the group exists, it has not been consulted, and nothing in the sections above should be read as a description of what was actually said in it.
• No numbers. There are no subscriber counts, no article counts, no readership estimates and no measure of how widely the group propagated. The two files that name the group carry no traffic information at all: the article-number fields in the active file are identical placeholders on every line.
• No standing documents of its own. No FAQ, welcome message or periodic posting for alt.newbie was found in the archives consulted. This is entirely normal for an alt.* group and is not evidence that none existed; the *.answers archive collected documents that were submitted to it, and most alt.* groups never submitted anything.
• No account of the group’s actual temper. Whether the housewarming line described the group or merely named an aspiration is not something the surviving paperwork can answer. Who wrote the description, or when, is not recorded either: none of the four archived control messages for the group contains it, and no source consulted shows it being changed afterwards. The group had thirty-four years in which to diverge from it in either direction.
• No resolution of the founding joke. The 1992 control message proposes a forum for discussion of newbies and does it in a register that expects to be laughed at. The group’s subsequent reputation is as a place beginners could ask questions. Both are on the record. The transition between them is not.
• No individuals. The names appearing in this article are those attached to public control messages and published documents. No moderator, regular or volunteer of alt.newbie is named, because none is documented.

What the record does show is unusually complete for an alt.* group of no fame: the exact wording of its creation, its charter, its removal, its restoration, the removal of its near-duplicate, the line by which it has been described in the reference file for three decades, and the fact that both it and its sibling remain in the canonical list today.

Scope and limits

This is a directory page, not a history of Usenet, and its boundaries are worth stating. news2mail.com was a mail-to-news gateway that operated between 2000 and 2004; this directory preserves the groups that gateway carried, and alt.newbie is here because it was one of them. The gateway’s snapshot has no bearing on the group’s importance, only on which groups this site happens to document.

The group-specific facts above come from two public archives mirrored by the Internet Systems Consortium: the active and newsgroups files, consulted on 26 August 2026, and the collection of newgroup and rmgroup control messages seen on the network, which is comprehensive from early 1991 onward with a known gap of about a month around August 1992. Control-message archives record what reached the collecting sites, which is not identical to everything ever sent; absence from them is weak evidence.

The netiquette material is quoted from RFC 1855 and from Brad Templeton’s Emily Postnews column as published by its author; the lexical material from the Jargon File; the account of the September effect from standard reference sources, with the one dating discrepancy between them noted where it arises rather than resolved. The successor platforms described in the penultimate section postdate the gateway: phpBB’s first release is contemporary with the gateway’s early years, but Stack Overflow, the instrumented platforms that followed it, and the retirement of the large web gateway to Usenet in 2024 all fall after 2004 and are included as context for what happened to the function, not as part of the group’s own story.

Adjacent material is deliberately not duplicated, and the pages that own it are linked from the sections above rather than listed again here: the hierarchy’s creation mechanics and its temperament belong to the alt.* hub, the administrative groups and the official induction machinery to the news.* hub, the answers family and the FAQ as a genre to the soc.answers page, and the medium itself to this site’s introduction to Usenet. Where a claim on this page rests on one of those, that page is the citation.