| < Photos | Russ Allbery | NNTP > |
Usenet interprets management as damage and routes around it.
Peter da Silva
The still-current standard for the format of Usenet articles is RFC 1036. It is unfortunately quite old and ambiguous in places, and refers to RFC 822 instead of the current standard for mail messages (RFC 5322). It's probably the best guide for implementors but must be taken with a grain of salt and doesn't fully document the current state of Usenet.
One widely referred-to later draft is the so-called Son of 1036, an old Internet Draft that was intended to be a replacement for 1036 but was never formally published. It's more accurate and comprehensive in places, less accurate in others, and documents some changes that were never deployed and probably never will. Anyone implementing Usenet software has to check it closely against existing practice before relying on it.
Work has recently been completed in the now-closed IETF USEFOR working group on writing a new standard for the Usenet article format that reflects existing practice and incorporates MIME. For more information, see the USEFOR pages. This is also the best collection of links related to Usenet standards that I know of.
Usenet (netnews) articles use the same basic header format as e-mail messages and HTTP headers and share the same IANA registry for header field names. Most of the references here for netnews header field names are to the current USEFOR drafts.
The work of the USEFOR working group was divided into three separate documents: USEFOR, which specifies the basic article format; USEPRO, which specifies the high-level protocol followed by Netnews servers and clients when manipulating messages; and USEAGE (sic), a best-practices guide to implementing news software.
USEFOR and USEPRO are now in AUTH48, meaning that they're in the final stage of revision before publication as RFCs.
I am the current document editor for the USEPRO document. You can see the latest released version of my draft here:
The following drafts have been prepared under the aegis of the USEFOR working group. Not all of these are being currently pursued.
If you're curious about the history of these documents, or want to see the other drafts that have gone into the working group discussions, you can review my draft archive. I may have missed a few, but most of them should be here. Also see the IETF Internet-Draft search for USEFOR working group drafts.
This is a set of best-practice guidelines for netnews moderators written back in 1994. This was intended to be published as an RFC, but was never completed. Some of the advice is out of date, but much of this information is still relevant.
The PGP Moose protocol specifies a mechanism for signing articles including certain key headers so that the resulting signature can be used to check several key header fields and the newsgroups to which the article was posted. This protocol is used primarily to validate approvals to moderated groups.
This document is the original README by Greg Rose that accompanied the reference implementation of PGP Moose. The canonical version is on Greg Rose's web site.
The pgpverify protocol is another protocol for signing a Usenet article that includes a different set of metadata and header fields than the PGP Moose protocol. It's used primarily (and very widely) to authenticate control messages.
This document is the FORMAT document for the pgpcontrol software. The canonical version is in the pgpcontrol distribution site.
This was the original Usenet standard. It documents the so-called "B News" format, the format implemented by the B news software. It includes both the article format standard and the original standard for transmission of articles using UUCP or e-mail, including rnews batching. It was obsoleted by RFC 1036 and is now of only historical interest.
| < Photos | Russ Allbery | NNTP > |