Summary

  • Newsgroups identifies where the current article is posted; Followup-To requests where a newly composed followup should go. The precursor remains visible where it was filed.
  • The special lowercase value poster changes the default medium: the response should be emailed to Reply-To, or to From when no reply address exists, rather than posted to a newsgroup.
  • RFC 5537 treats both forms as posting-agent defaults that the new poster may override. References preserves ancestry and Distribution preserves scope separately, so destination is neither lineage nor access control.

One article in two rooms, one reply in a third

Imagine a notice crossposted to two technical groups. Readers in both can see the same precursor. Its Followup-To names a third group devoted to discussion. When a reader chooses “follow up,” the client prepares a new article for that third room.

Nothing has been recalled from the first two groups. Their readers have not lost access, and the original article has not changed location. Only the destination proposed for a future article is different.

This was a compact answer to a recurring problem. Crossposting could make an announcement visible to several communities, but copying every response back into all of them would multiply traffic and fragment attention. A portable reply default could let discovery remain broad while conversation converged.

The first separation was already explicit

RFC 850 gave Followup-To the same group-list form as Newsgroups. When present, followup articles were to be posted to its listed groups. When absent, they normally returned to the precursor's groups, with an early special case that diverted replies from the broad net.general group into net.followup.

The old exception is historically revealing. Even before a general field was widely used, the network recognized that the best place to read an announcement need not be the best place to conduct its discussion.

The field did not rewrite the current Newsgroups value. It influenced the construction of the next article. That distinction kept the precursor's publication record stable while allowing a conversation to acquire a different operating venue.

A single word could change the medium

RFC 1036 preserved group redirection and introduced the special keyword poster. In that case, followups were not to be posted; the response should instead be mailed to the submitter.

RFC 5536 makes the request more precise. Followup-To specifies where the poster has requested that followups be posted and should normally be omitted when it merely repeats Newsgroups. Absence already carries the ordinary return-to-source default.

The lowercase token poster is exceptional syntax, not a newsgroup name. It asks the client to email the response directly, using Reply-To when present and otherwise From. The original article remains public in its newsgroups. Only the default action for the new response crosses from Netnews into mail.

Mail brought its own addressing authority

The fallback did not invent email semantics inside Netnews. RFC 5322 separately defines Reply-To as the address or addresses to which the author suggests replies be sent. If that field is absent, replies should normally use From, unless the person composing the reply chooses otherwise.

This layering matters. Followup-To: poster chooses a medium and points to the article's reply-address evidence. Reply-To or From supplies the mailbox destination. Mail transport then decides whether delivery succeeds. No single field proves mailbox ownership, reachability or receipt.

Nor does the private default make the precursor confidential. Everyone who could read the original can still do so. A reader may quote it elsewhere, and the next poster may override the default. The token organizes expected behaviour; it does not create a privacy boundary.

The next poster still owned the new publication

RFC 5537 describes followup construction as a set of defaults. If the precursor says poster, the posting agent must not post by default and instead prepares an email reply. The user may override that choice, after which the client proceeds as though the field were absent.

For a group-list value, the new article's Newsgroups should inherit the precursor's Followup-To. If the field is absent, it inherits the precursor's Newsgroups. Again, the poster may change the initial content before submission.

The allocation of authority is deliberate. The first poster can request a useful place for discussion. The reading agent can present that choice safely. The person writing the response remains responsible for the audience of a new publication. A header in an earlier article cannot permanently legislate every descendant.

Destination, ancestry and scope stayed independent

A followup must also carry References linking it to its precursor. That lineage survives even when the new article is posted to a different group. Moving the discussion does not erase where it came from.

Distribution is inherited through a separate rule. A new group destination does not automatically grant wider propagation, and a narrow distribution does not decide which group receives the reply. Subject, audience, ancestry and scope are related, but no one field is allowed to impersonate all four.

That separation also exposes abuse. RFC 5537 warns that requesting followups to unrelated groups can disrupt those groups. A valid name is not proof that the destination is topically legitimate. Clients and users need to see the transition before sending rather than silently obeying an inherited value.

Registration stabilized a request, not a route

The current IANA Message Headers Registry lists Followup-To as a standard Netnews field referenced to RFC 5536. Implementations can agree on the name and grammar.

The registry does not establish present-day client support, prove that a user reviewed the destination, authenticate a reply mailbox or guarantee posting or mail delivery. The official sources define protocol history and defaults, not current adoption.

The historical achievement was to make a conversational handoff visible without pretending it was compulsory. The article could stay where it had been read. The response could begin somewhere more appropriate. The boundary between them was a new author's decision, assisted—but not owned—by the old article.