Summary

  • RFC 1855 was an Informational, deliberately adaptable minimum. Its RFC number preserved guidance; it did not make the guidance an Internet Standard or give the IETF an enforcement arm.
  • The guide separated one-to-one communication, one-to-many publication and information services, then assigned different duties to users, administrators, moderators and service operators.
  • Its most durable lesson is evidentiary: route a complaint to the actor who can investigate it, distinguish personal speech from organizational authority, and do not turn a convention, header or accusation into proof.

A rulebook that began by narrowing itself

The memorable details of RFC 1855 invite nostalgia. Do not type whole messages in capitals. Keep a signature short because someone may be paying for the connection by the minute. Read a group before posting. Wait overnight before sending an angry answer. Do not mistake a smiley for a licence to insult someone.

Published in October 1995 as FYI 28, the document put its most important sentence before any of that advice. It did not specify an Internet standard. Its abstract called it a minimum set of guidelines that organizations might take and adapt. The bullets were chosen to make particular guidance easier to retrieve as well as to adapt.

This was a precise response to the problem the authors described. The Internet’s population was widening beyond people who had grown up with its transport and protocols. New participants did not need to become network engineers before speaking. They did need a way to discover what a channel did to their words, whose resources it consumed and which rules governed the account through which they entered.

The answer was not a world etiquette police. The guide told a university student, corporate employee or ISP customer to examine the rules of the organization supplying access and consult local authority for specifics. A shared publication could orient conduct; the enforceable boundary still had to be attached to a real service, organization and decision maker.

Three rooms, not one public square

RFC 1855 divided the network into three social geometries. One-to-one mail and talk resembled a conversation. One-to-many mailing lists and NetNews could expose one act to a distributed audience. Information services—FTP, the Web, WAIS, Gopher, MUDs and MOOs—placed users inside resources operated by someone else.

The same keystroke acquired different consequences in each room. A private answer might interrupt one person. A list reply could reach hundreds. A NetNews article could propagate between hosts and arrive after its reply. A file retrieval could impose load on a remote service. Etiquette was therefore not merely tone. It was knowledge of fan-out, persistence, cost and custody.

This is why the guide repeatedly worried about address and audience. A group address could look like a person. Reply machinery could send an intended private response to an entire list. A NetNews Distribution header might try to limit reach, but the document warned that the delivery system made the field unreliable. The sender was asked to understand the path rather than treat the interface’s label as a guarantee.

The asymmetry became sharp after transmission. Mailing-list messages generally could not be retrieved even by the sender’s administrator. News and list archives could preserve words for a long time. The send action was local; its downstream copies were not. Responsibility therefore began before the click, at the last point where the author still controlled the text and audience choice.

A map of offices

The guide did not leave every problem with the user. It named distinct operational offices.

Users were responsible for address selection, subject lines, quotation, disclosure, audience fit and the costs their messages imposed. They were told to learn a list’s subscription mechanism and send administrative requests to the right address. A misplaced unsubscribe was not just embarrassing; it demonstrated that content and control traffic had different destinations.

Administrators were told to publish site rules, archive policy and logging policy; monitor system health; maintain working aliases such as Postmaster; and respond to complaints. They were also told to investigate accusations with an open mind because addresses could be forged. The administrator had a duty to examine evidence, not a licence to convert the first complaint into a verdict.

Moderators had a narrower surface. They were to keep FAQs, welcome messages, subscription instructions and group charters available, process material in time and arrange a substitute when absent. Their authority came from maintaining a particular forum. It did not travel automatically to another list, another service or the Internet as a whole.

Information-service operators occupied another role. The guide said the people paying for a service could make its usage rules, and it asked them to explain what could be copied, maintain documentation, disclose data-use intentions and test across different clients. That is service authority tied to responsibility for a working surface, not ownership of every user or idea that crossed it.

A voice did not become a constituency

One short instruction carries unusual institutional weight: assume that individuals speak for themselves and do not represent their organizations unless they say so explicitly. The network made circulation cheap; it did not manufacture delegation.

A message from a company domain could be a personal opinion. A participant in a working group could report experience without binding an employer. A moderator could shape a forum without speaking for every reader. RFC publication itself could preserve a working group’s guidance without turning every item into a mandatory rule.

The distinction protects participation. People can contribute warnings, experience and criticism without every sentence being mistaken for an institutional act. It also protects institutions from invented consent. Attendance, access, speech, moderation and authorization are related, but they are not interchangeable records.

A complaint was not identity proof

RFC 1855’s advice about forgery prevents its social guidance from becoming a shortcut around evidence. Mail and news could be spoofed to different degrees. Administrators were told to apply reality checks and to investigate complaints without assuming the apparent address settled who acted.

That boundary matters because abuse response has two possible victims. One is the recipient of the unwanted or harmful message. The other is the innocent user, site or administrator to whom an unreliable artifact points. Fast escalation can protect the first while injuring the second if provenance is weak.

The document did not supply modern authentication, a universal evidence format or an appeal court. It supplied a discipline: retain the accusation, identify the relevant local operator, test the available records and distinguish the ability to act from proof that the proposed target is responsible.

Courtesy had an economic substrate

Some recommendations look quaint because the numbers changed. A four-line signature, a 50-kilobyte attachment threshold and a short terminal line belonged to the cost and display constraints of the time. Their underlying mechanism has not aged as quickly.

RFC 1855 observed that sender and recipient, or their organizations, shared the cost of delivering mail through bandwidth, disk and CPU use. Unsolicited volume allowed one party to spend another party’s resources. A message could be cheap for the sender and collectively expensive after replication.

One-to-many etiquette therefore governed externalities. Asking whether the audience wanted the content was also asking who had authorized the use of their scarce capacity and attention. The guide did not calculate a universal acceptable price. It located the burden: the person choosing fan-out had to consider costs that the send button did not display.

Publication did not certify truth

The information-services section extended the same restraint from people to documents. A familiar filename extension might not describe the file. A README could be stale. A service naming convention might not be enforced. Material found online was not necessarily current or accurate merely because new publishing tools made it reachable.

Operators, in turn, were told to maintain their information, date time-sensitive material, state policies and disclose what they planned to do with submitted feedback. Publishing created a maintenance obligation; it did not make the publisher an oracle.

That distinction binds the whole guide together. A convention is a clue, a message is a claim, a complaint is a trigger, and an archive is evidence of publication. None becomes a complete decision merely by appearing on the network.

Spam made the responsibility map harder

Four years later, RFC 2635 returned to mass unsolicited mail and postings. It was also Informational. Its response was again divided among users, system and news administrators, mailing-list managers and providers.

The newer document described why unwanted messages multiplied costs and why careless replies could amplify them. It also warned that countermeasures could block legitimate mail, complaints could overwhelm or punish the wrong postmaster, and escalation had to follow local procedure. The existence of harmful conduct did not erase proportionality or provenance.

This was the operational maturation of netiquette. A social expectation could lead to filters, account terms and provider action, but each mechanism brought a new accountable owner and a new failure mode. “The community disapproves” was not enough to explain who filtered a message, under which rule, with which evidence or how a false positive would be repaired.

An RFC number was not a crown

RFC 2026, published the following year, described the RFC series as the official publication channel for standards documents and other Internet-community publications. It distinguished the STD and BCP subseries from Informational and Experimental material outside the standards track. RFC 8729 later restated the broader archival purpose: the series contains standards, but also research, engineering thought and other contributions.

That history matters because “it is an RFC” is not a complete statement about authority. The number tells us that a document entered a durable publication series. Its stream, category, status and actual control surface tell us what kind of claim it can make.

RFC 1855 was valuable without being law. It compressed accumulated practice into a findable minimum, labelled its limits and directed readers toward the local rules and people capable of acting. Its restraint is the point. The early Internet did not need one global referee in order to make responsibility more legible.

What the record cannot prove

The four documents do not measure how many people read or adopted the guidelines. They do not establish present platform rules, compliance rates or the success of any sanction. Interface-specific numbers from 1995 are historical evidence, not current universal thresholds.

Nor should the forceful language inside the guide be smoothed away. It sometimes predicted revoked connectivity or described conduct as forbidden. Those phrases document expectations in a particular operating culture. They do not cancel the explicit Informational status, identify a single enforcement institution or prove that every service applied the same rule.

The defensible conclusion is narrower and stronger. The etiquette RFC made behavior auditable by naming audience, medium, local policy and responsible role. It showed how a distributed network could publish common guidance while leaving execution attached to bounded operators. The rule became useful because the referee remained local and identifiable.

Sources