INTERNET-DRAFT Charles H. Lindsey Usenet Format Working Group University of Manchester September 2004 News Article Architecture and Protocols Status of this Memo By submitting this Internet-Draft, I certify that any applicable patent or other IPR claims of which I am aware have been disclosed, and any of which I become aware will be disclosed, in accordance with RFC 3668. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. This Internet-Draft will expire in March 2005. Abstract This Draft, together with its companion draft [USEFOR], are intended as standards track documents, together obsoleting RFC 1036, which itself dates from 1987. This Standard defines the architecture of Netnews systems and specifies the requirements to be met by software which originates, distributes, stores and displays Netnews articles. Since the 1980s, Usenet has grown explosively, and many Internet and non-Internet sites now participate. In addition, the Netnews technology is now in widespread use for other purposes. Backward compatibility has been a major goal of this endeavour, but where this standard and earlier documents or practices conflict, this standard should be followed. In most such cases, current practice is already compatible with these changes. A companion Current Best Practice document [USEAGE], addressing requirements which are present for Social rather than Normative C. H. Lindsey [Page 1] News Article Architecture and Protocols September 2004 reasons is in preparation. [The use of the words "this standard" within this document when referring to itself does not imply that this draft yet has pretensions to be a standard, but rather indicates what will become the case if and when it is accepted as an RFC with the status of a proposed or draft standard.] [Remarks enclosed in square brackets and aligned with the left margin, such as this one, are not part of this draft, but are editorial notes to explain matters amongst ourselves, or to point out alternatives, or to assist the RFC Editor.] [In this draft, references to [NNTP] are to be replaced by references to the RFC arising from the series of drafts draft-ietf-nntpext-base-*.txt, which has now passed its IETF last call.] Table of Contents 1. Introduction .................................................. 4 1.1. Basic Concepts ............................................ 4 1.2. Objectives ................................................ 5 1.3. Historical Outline ........................................ 5 2. Definitions, Notations and Conventions ........................ 6 2.1. Definitions ............................................... 6 2.2. Defining the Architecture ................................. 7 2.3. Textual Notations ......................................... 8 3. Changes to the existing protocols ............................. 9 3.1. Principal Changes ......................................... 9 3.2. Transitional Arrangements ................................. 10 4. Transport ..................................................... 11 5. Definition of new Media Types ................................. 11 5.1. Application/news-transmission ............................. 11 5.2. Message/news obsoleted .................................... 12 5.3. Application/news-groupinfo ................................ 13 5.4. Application/news-checkgroups .............................. 14 6. Control Messages .............................................. 14 6.1. Digital Signature of Headers .............................. 15 6.2. Group Control Messages .................................... 15 6.2.1. The 'newgroup' Control Message ........................ 15 6.2.1.1. The Body of the 'newgroup' Control Message ........ 16 6.2.1.2. Initial Articles .................................. 16 6.2.1.3. Example ........................................... 17 6.2.2. The 'rmgroup' Control Message ......................... 17 6.2.2.1. Example ........................................... 18 6.2.3. The 'mvgroup' Control Message ......................... 18 6.2.3.1. Example ........................................... 19 6.2.4. The 'checkgroups' Control Message ..................... 20 6.3. Cancel .................................................... 21 6.4. Ihave, sendme ............................................. 22 6.5. Obsolete control messages. ............................... 24 7. Duties of Various Agents ...................................... 24 7.1. General principles to be followed ......................... 24 C. H. Lindsey [Page 2] News Article Architecture and Protocols September 2004 7.2. Duties of an Injecting Agent .............................. 25 7.2.1. Proto-articles ........................................ 25 7.2.2. Procedure to be followed by Injecting Agents .......... 26 7.2.3. Procedure for Forwarding to a Moderator ............... 28 7.3. Duties of a Relaying Agent ................................ 29 7.3.1. Path-Header Example ................................... 31 7.4. Duties of a Serving Agent ................................. 32 7.5. Duties of a Posting Agent ................................. 33 7.6. Duties of a Followup Agent ................................ 34 7.7. Duties of a Reading Agent ................................. 35 7.8. Duties of a Moderator ..................................... 35 7.9. Duties of a Gateway ....................................... 37 7.9.1. Duties of an Outgoing Gateway ......................... 38 7.9.2. Duties of an Incoming Gateway ......................... 39 7.9.3. Example ............................................... 41 8. Security and Related Considerations ........................... 42 8.1. Leakage ................................................... 42 8.2. Attacks ................................................... 42 8.2.1. Denial of Service ..................................... 42 8.2.2. Compromise of System Integrity ........................ 44 8.3. Liability ................................................. 45 9. IANA Considerations ........................................... 45 10. References ................................................... 45 11. Acknowledgements ............................................. 47 12. Contact Address .............................................. 47 Appendix A.1 - A-News Article Format .............................. 47 Appendix A.2 - Early B-News Article Format ........................ 48 Appendix A.3 - Obsolete Control Messages .......................... 48 Appendix B - Notices .............................................. 49 Appendix C - Change Log ........................................... 50 C. H. Lindsey [Page 3] News Article Architecture and Protocols September 2004 [This draft [USEPRO] and its partner [USEFOR] are an interim stage in the splitting into two parts of the earlier draft [ARTICLE]. There is a certain amount of material - basic concepts, definitions, etc - which ultimately need occur in only one of the documents, and further such material which may not be needed at all (e.g. terms currently defined which in the event may not get used). For the moment, all such material has been retained in the present draft (it being, in any case, easier to take unwanted stuff out than to put new stuff in). It has also to be decided, for such material which is needed by both documents, which one (the "Primary") should contain it and which one should incorporate it by reference (essentially, this draft is written so that it could be the Primary).] [Again, references in this document to sections which ultimately are expected to be in [USEFOR] are still to the old sections in [ARTICLE], since [USEFOR] is not stabilized enough to refer to specific sections in it yet. These references can be recognized by being prefixed with "a-". These references can be fixed up later as we find the proper places for them to point to. In any case, there will doubtless be more toing and froing of texts between the two documents before we are done.] 1. Introduction 1.1. Basic Concepts "Netnews" is a set of protocols for generating, storing and retrieving news "articles" (which resemble email messages) and for exchanging them amongst a readership which is potentially widely distributed. It is organized around "newsgroups", with the expectation that each reader will be able to see all articles posted to each newsgroup in which he participates. These protocols most commonly use a flooding algorithm which propagates copies throughout a network of participating servers. Typically, only one copy is stored per server, and each server makes it available on demand to readers able to access that server. An important characteristic of Netnews is the lack of any requirement for a central administration or for the establishment of any controlling host to manage the network. A set of hosts within a network which, by mutual arrangement, operates some variant (whether more or less restrictive) of the Netnews protocols is a "cooperating subnet". "Usenet" is a particular worldwide publicly accessible network based upon the Netnews protocols, with the newsgroups being organized into recognized "hierarchies". Anybody can join (it is simply necessary to negotiate an exchange of articles with one or more other participating hosts). A "policy" is a rule intended to facilitate the smooth operation of a network by establishing parameters which restrict behaviour that, whilst technically unexceptionable, would nevertheless contravene some accepted standard of "Good Netkeeping". Since the ultimate beneficiaries of a network are its human readers, who will be less C. H. Lindsey [Page 4] News Article Architecture and Protocols September 2004 tolerant of poorly designed interfaces than mere computers, articles in breach of established policy can cause considerable annoyance to their recipients. 1.2. Objectives The purpose of this present standard is to define the overall architecture and the protocols to be used for Netnews in general, and for Usenet in particular, and to set standards to be followed by software that implements those protocols. A companion standard [USEFOR] sets out the canonical format of news articles exchanged between the various agents comprising that architecture. It is NOT the purpose of this standard to settle matters of policy, nor aspects of software behaviour which do not impinge upon the generation, transmission, storage and reception of articles, nor how the authority of various agencies to create such policies and to exercise control or oversight of the various parts of Usenet is established. For these purposes, a separate Best Current Practice document [USEAGE] is being provided. Nevertheless, it is assumed that such agencies with the necessary authority will exist, and tools are provided within the protocols for their use. 1.3. Historical Outline Network news originated as the medium of communication for Usenet, circa 1980. Since then, Usenet has grown explosively, and many Internet and non-Internet sites participate in it. In addition, the news technology is now in widespread use for other purposes, on the Internet and elsewhere. The earliest news interchange used the so-called "A News" article format. Shortly thereafter, an article format vaguely resembling Internet Mail was devised and used briefly. Both of those formats are completely obsolete; they are documented in Appendix A.1 and Appendix A.2 for historical reasons only. With publication of [RFC 850] in 1983, news articles came to closely resemble Internet Mail messages, with some restrictions and some additional headers. [RFC 1036] in 1987 updated [RFC 850] without making major changes. A Draft popularly referred to as "Son of 1036" [Son-of-1036] was written in 1994 by Henry Spencer. Much is taken directly from Son of 1036, and it is hoped that we have followed its spirit and intentions. [It is anticipated that [Son-of-1036] will shortly be published as an informational RFC (for purposes of historical documentation only), in which case most historical information can be removed from this draft, including the whole of Appendix A.1 and Appendix A.2.] C. H. Lindsey [Page 5] News Article Architecture and Protocols September 2004 2. Definitions, Notations and Conventions 2.1. Definitions An "article" is the unit of news, synonymous with an [RFC 2822] "message". A "proto-article" (7.2.1) is one that has been created by a "posting agent" but has not yet been injected into a Netnews system system by an "injecting agent". It may lack some otherwise mandatory headers. A "message identifier" (a-5.3) is a unique identifier for an article, usually supplied by the posting agent which posted it or, failing that, by the injecting agent. It distinguishes the article from every other article ever posted anywhere. Articles with the same message identifier are treated as if they are the same article regardless of any differences in the body or headers. A "newsgroup" is a single news forum, a logical bulletin board, having a name and nominally intended for articles on a specific topic. An article is "posted to" a single newsgroup or several newsgroups. When an article is posted to more than one newsgroup, it is said to be "crossposted"; note that this differs from posting the same text as part of each of several articles, one per newsgroup. A newsgroup may be "moderated", in which case submissions are not posted directly, but mailed to a "moderator" for consideration and possible posting. Moderators are typically human but may be implemented partially or entirely in software. A "hierarchy" is the set of all newsgroups whose names share a first component (as defined in a-5.5). The term "sub-hierarchy" is also used where several initial components are shared. A "poster" is the person or software that composes and submits a possibly compliant article to a posting agent. The poster is synonymous with [RFC 2822]'s author. A "reader" is the person or software reading news articles. A "followup" is an article containing a response to the contents of an earlier article (the followup's "precursor"). [Alternative definition, to bu ised if similar woprding is not added to the description of the References-header (a-6.10):] A "followup" is an article containing a response to the contents of an earlier article (its "precursor"), or which is otherwise intended to be grouped with that article for purposes of display (e.g. as part of a multipart posting such as a FAQ). An (email) "address" is the mailbox [RFC 2822] (or more particularly the addr-spec within that mailbox) which directs the delivery of an email to its intended recipient, who is said to "own" that address. C. H. Lindsey [Page 6] News Article Architecture and Protocols September 2004 An article's "reply address" is the address to which mailed replies should be sent. This is the address specified in the article's From- header (a-5.2), unless it also has a Reply-To-header (a-6.1). A "sender" is the person or software (usually, but not always, the same as the poster) responsible for the operation of the posting agent or, which amounts to the same thing, for passing the article to the injecting agent. A "control message" is an article which is marked as containing control information; a "serving agent" receiving such an article may (subject to the policies observed at that site) take actions beyond just filing and passing on the article. 2.2. Defining the Architecture A Netnews system is a distributed database composed of "agents" of various types which, acting together according to the protocols defined in section 7 of this standard, causes articles to be propagated throughout the system and to be made available to its readers. The protocols ensure that all copies of a given article, wherever stored, are identical apart from those headers defined as variant (a-4.2.5.3). A "posting agent" is the software that assists posters to prepare proto-articles in compliance with [USEFOR]. The proto-article is then passed on to an "injecting agent" for final checking and injection into the news stream. If the article is not compliant, or is rejected by the injecting agent, then the posting agent informs the poster with an explanation of the error. A "reading agent" is software which presents articles to a reader. A "followup agent" is a combination of reading agent and posting agent that aids in the preparation and posting of a followup. [Alternative definition, to be used if the alternative definition for "followup" is used:] A "followup agent" is a combination of reading agent and posting agent that aids in the preparation and posting of a response intended as a followup to a precursor. An "injecting agent" takes the finished article from the posting agent (often via the NNTP "POST" command), performs some final checks and passes it on to a "relaying agent" for general distribution. It is expected to bear some responsibility towards the rest of the network for the behaviour of its posters (and provision is therefore made for it to be easily contactable by email). A "relaying agent" is software which receives allegedly compliant articles from injecting agents and/or other relaying agents, and possibly passes copies on to other relaying agents and "serving agents". C. H. Lindsey [Page 7] News Article Architecture and Protocols September 2004 A "serving agent" receives an article from a relaying agent and files it in a "news database". It also provides an interface for reading agents to access the news database. A "news database" is the set of articles and related structural information stored by a serving agent and made available for access by reading agents. A "gateway" is software which receives news articles and converts them to messages of some other kind (e.g. mail to a mailing list), or vice versa; in essence it is a translating relaying agent that straddles boundaries between different methods of message exchange. The most common type of gateway connects newsgroup(s) to mailing list(s), either unidirectionally or bidirectionally, but there are also gateways between news networks using this standard's news format and those using other formats. Posting, reading and followup agents (which are usually just different services provided by the same piece of software) are known collectively as "user agents". 2.3. Textual Notations This standard contains explanatory NOTEs using the following format. These may be skipped by persons interested solely in the content of the specification. The purpose of the notes is to explain why choices were made, to place them in context, or to suggest possible implementation techniques. NOTE: While such explanatory notes may seem superfluous in principle, they often help the less-than-omniscient reader grasp the purpose of the specification and the constraints involved. Given the limitations of natural language for descriptive purposes, this improves the probability that implementors and users will understand the true intent of the specification in cases where the wording is not entirely clear. "US-ASCII" is short for "the ANSI X3.4 character set" [ANSI X3.4]. US-ASCII is a 7 bit character set. Please note that this standard requires that all agents be 8 bit clean; that is, they must accept and transmit data without changing or omitting the 8th bit. Certain words, when capitalized, are used to define the significance of individual requirements. The key words "MUST", "REQUIRED", "SHOULD", "RECOMMENDED", "MAY" and "OPTIONAL", and any of those words associated with the word "NOT", are to be interpreted as described in [RFC 2119]. NOTE: The use of "MUST" or "SHOULD" implies a requirement that would or could lead to interoperability problems if not followed. C. H. Lindsey [Page 8] News Article Architecture and Protocols September 2004 NOTE: A requirement imposed on a relaying or serving agent regarding some particular article should be understood as applying only if that article is actually accepted for processing (since any agent may always reject any article entirely, for reasons of site policy). Wherever the context permits, use of the masculine includes the feminine and use of the singular includes the plural, and vice versa. Throughout this standard we will give examples of various definitions, headers and other specifications. It needs to be remembered that these samples are for the aid of the reader only and do NOT define any specification themselves. In order to prevent possible conflict with "Real World" entities and people the top level domain ".example" is used in all sample domains and addresses. The hierarchy "example.*" is also used as a sample hierarchy. Information on the ".example" top level domain is in [RFC 2606]. 3. Changes to the existing protocols This standard prescribes many changes, clarifications and new features since the protocols described in [RFC 1036] and [Son-of- 1036]. It is the intention that they can be assimilated into Usenet as it presently operates without major interruption to the service, though some of the new features may not begin to show benefit until they become widely implemented. This section summarizes the main changes, and comments on some features of the transition. 3.1. Principal Changes o The [RFC 2822] conventions for parenthesis-enclosed comments in headers are supported. o Whitespace is permitted in Newsgroups-headers, permitting folding of such headers. Indeed, all headers can now be folded. o An enhanced syntax for the Path-header enables the injection point of and the route taken by an article to be determined with certainty. o Large parts of MIME are recognized as an integral part of Netnews. o There is a new Control message 'mvgroup' to facilitate moving a group to a different place (name) in a hierarchy. o There is a new mandatory Injection-Date-header to facilitate the rejection of stale articles. o There are several new optional headers defined, notably Archive, Complaints-To, Injection-Info, Mail-Copies-To, Posted-And-Mailed and User-Agent, leading to increased functionality. o Provision has been made for almost all headers to have MIME-style parameters (to be ignored if not recognized), thus facilitating extension of those headers in future standards. o Certain headers and Control messages (a-Appendix A.3 and Appendix A.3) have been made obsolete. o Distributions are expected to be checked at the receiving end, as well as the sending end, of a relaying link. o There are numerous other small changes, clarifications and C. H. Lindsey [Page 9] News Article Architecture and Protocols September 2004 enhancements. 3.2. Transitional Arrangements An important distinction must be made between serving and relaying agents, which are responsible for the distribution and storage of news articles, and user agents, which are responsible for interactions with users. It is important that the former should be upgraded to conform to this standard as soon as possible to provide the benefit of the enhanced facilities. Fortunately, the number of distinct implementations of such agents is rather small, at least so far as the main "backbone" of Usenet is concerned, and many of the new features are already supported. Contrariwise, there are a great number of implementations of user agents, installed on a vastly greater number of small sites. Therefore, the new functionality has been designed so that existing user agents may continue to be used, although the full benefits may not be realised until a substantial proportion of them have been upgraded. In the list which follows, care has been taken to distinguish the implications for both kinds of agent. o [RFC 2822] style comments in headers do not affect serving and relaying agents (note that the Message-ID-, Newsgroups-, Distribution- and Path-headers do not contain them). They are unlikely to hinder their proper display in existing reading agents except in the case of the References-header in agents which thread articles. Therefore, it is provided that they SHOULD NOT be generated except where permitted by the previous standards. o Because of its importance to all serving agents, the extension permitting whitespace and folding in Newsgroups-headers SHOULD NOT be used until it has been widely deployed amongst relaying agents. User agents are unaffected. o The new style of Path-header is already consistent with the previous standards. However, the intention is that relaying agents should eventually reject articles in the old style, and so this possibility should be offered as a configurable option in relaying agents. User agents are unaffected. o The introduction of MIME reflects a practice that is already widespread. Articles in strict compliance with the previous standards (using strict US-ASCII) will be unaffected. Many user agents already support it, at least to the extent of widely used charsets such as ISO-8859-1. Users expecting to read articles using other charsets will need to acquire suitable reading agents. It is not intended, in general, that any single user agent will be able to display every charset known to IANA, but all such agents MUST support US-ASCII. Serving and relaying agents are not affected. o The new Control: mvgroup command will need to be implemented in serving agents. For the benefit of older serving agents it is therefore RECOMMENDED that it be followed shortly by a corresponding newgroup command and it MUST always be followed by a rmgroup command for the old group after a reasonable overlap C. H. Lindsey [Page 10] News Article Architecture and Protocols September 2004 period. An implementation of the mvgroup command as an alias for the newgroup command would thus be minimally conforming. User agents are unaffected. o Provision is made for relaying and serving agents to use the Date-header in the case of articles injected through existing agents which do not provide an Injection-Date-header. o All the headers newly introduced by this standard can safely be ignored by existing software, albeit with loss of the new functionality. 4. Transport As in this standard's predecessors, the exact means used to transmit articles from one host to another is not specified. NNTP [NNTP] is the most common transmission method on the Internet, but much transmission takes place entirely independent of the Internet. Other methods in use include the UUCP protocol [RFC 976] extensively used in the early days of Usenet, FTP, downloading via satellite, tape archives, and physically delivered magnetic and optical media. Transmission paths for news articles MUST treat news articles as uninterpreted sequences of octets, excluding the values 0 (US-ASCII NUL) and 13 and 10 (US-ASCII CR and LF, which MUST ONLY appear in the combination CRLF which denotes a line separator). NOTE: this corresponds to the range of octets permitted for MIME "8bit data" [RFC 2045]. Thus raw binary data cannot be transmitted in an article body except by the use of a Content- Transfer-Encoding such as base64. In particular, transmission paths MUST convey all headers (including body part headers and headers within message/rfc822 objects) intact, even if they contain octets in the range 128 to 255. These requirements include the transmissiom paths between posting agents, injecting agents, relaying agents, serving agents and reading agents, but NOT the paths traversed by Netnews articles that have been gatewayed into Email (7.9.1). [At some point it will be necessary for the IMAP standards to catch up with these requirements.] 5. Definition of new Media Types This standard defines (or redefines) several new Media Types, which require to be registered with IANA as provided for in [RFC 2048]. 5.1. Application/news-transmission The Media Type "application/news-transmission" is intended for the encapsulation of complete news articles where the intention is that the recipient should then inject them into Netnews. This Application type provides one of the methods for mailing articles to moderators (see 7.2.2) and it is also the preferred method when sending to an email-to-news gateway (see 7.9.2). C. H. Lindsey [Page 11] News Article Architecture and Protocols September 2004 NOTE: The benefit of such encapsulation is that it removes possible conflict between news and email headers and it provides a convenient way of "tunnelling" a news article through a transport medium that does not support 8bit characters. The MIME Media Type definition of "application/news-transmission" is: MIME type name: application MIME subtype name: news-transmission Required parameters: none Optional parameters: usage=moderate usage=inject usage=relay Encoding considerations: A transfer-encoding (such as Quoted- Printable or Base64) different from that of the article transmitted MAY be supplied (perhaps en route) to ensure correct transmission over some 7bit transport medium. Security considerations: A news article may be a "control message", which could have effects on the recipient host's system beyond just storage of the article. However, such control messages also occur in normal news flow, so most hosts will already be suitably defended against undesired effects. Published specification: [USEPRO] Body part: A complete article or proto-article, ready for injection into Netnews, or a batch of such articles in the batch format described in section 6.4. NOTE: It is likely that the recipient of an "application/news- transmission" will be a specialized gateway (e.g. a moderator's submission address) able to accept articles with only one of the three usage parameters "moderate", "inject" and "relay", hence the reason why they are optional, being redundant in most situations. Nevertheless, they MAY be used to signify the originator's intention with regard to the transmission, so removing any possible doubt. When the parameter "relay" is used, or implied, the body part MAY be a batch of articles to be transmitted together, in which case the batch format defined in section 6.4 MUST be used. 5.2. Message/news obsoleted The Media Type "message/news", as previously registered with IANA, is hereby declared obsolete. It was never widely implemented, and its default treatment as "application/octet-stream" by agents that did not recognize it was counter productive. The Media Type "message/rfc822" SHOULD be used in its place. C. H. Lindsey [Page 12] News Article Architecture and Protocols September 2004 5.3. Application/news-groupinfo The "application/news-groupinfo" is used in conjunction with the "newgroup" (6.2.1) and "mvgroup" (6.2.3) control messages. The newsgroup-name in the newsgroups-line MUST agree with the newsgroup- name in the "newgroup" or "mvgroup" control message. The Media Type "application/news-groupinfo" MUST NOT be used except as a part of such control messages. The "application/news-groupinfo" body part contains brief information about a newsgroup, i.e. the group's name, it's newsgroup-description and the moderation-flag. NOTE: The presence of the newsgroups-tag "For your newsgroups file:" is intended to make the whole newgroup message compatible with current practice as described in [Son-of-1036]. The MIME Media Type definition of "application/news-groupinfo" is: MIME type name: application MIME subtype name: news-groupinfo Required parameters: none Disposition: by default, inline Encoding considerations: "7bit" or "8bit" is sufficient and MUST be used to maintain compatibility. Security considerations: this type MUST NOT be used except as part of a control message for the creation or modification of a Netnews newsgroup Published specification: [USEPRO] The content of the "application/news-groupinfo" body part is defined as: groupinfo-body = [ newsgroups-tag CRLF ] newsgroups-line CRLF newsgroups-tag = %x46.6F.72 SP %x79.6F.75.72 SP %x6E.65.77.73.67.72.6F.75.70.73 SP %x66.69.6C.65.3A ; case sensitive ; "For your newsgroups file:" newsgroups-line = newsgroup-name [ 1*HTAB newsgroup-description ] [ 1*WSP moderation-flag ] newsgroup-description = utext *( *WSP utext ) moderation-flag = %x28.4D.6F.64.65.72.61.74.65.64.29 ; case sensitive "(Moderated)" The newsgroup-description MUST NOT contain any occurrence of the string "(Moderated)" within it. Although optional, the newsgroups- tag SHOULD be included until such time as this standard has been widely adopted, to ensure compatibility with present practice. C. H. Lindsey [Page 13] News Article Architecture and Protocols September 2004 Moderated newsgroups MUST be marked by appending the case sensitive text " (Moderated)" at the end. It is NOT recommended that the moderator's email address be included in the newsgroup-description as has sometimes been done. NOTE: There is no provision for the use of charsets other than US-ASCII within a newsgroup-description. Such a facility may be provided in a future extension to this standard. [That may seem harsh, but if we make any such provision now, it will make life more complicated and restrict our freedom when it comes to the proposed I18N extension. Therefore I resisted the temptation to include any charset parameter with this Media Type. Note that this also applies to the checkgroups message further on.] 5.4. Application/news-checkgroups The "application/news-checkgroups" Media Type is used in conjunction with the "checkgroups" control message (6.2.4). It MUST NOT be used except as a part of such control messages. The "application/news-checkgroups" body part contains a complete list of all the newsgroups in a (sub)hierarchy, their newsgroup- descriptions and their moderation status. The MIME Media Type definition of "application/news-checkgroups" is: MIME type name: application MIME subtype name: news-checkgroups Required parameters: none Disposition: by default, inline Encoding considerations: "7bit" or "8bit" is sufficient and MUST be used to maintain compatibility. Security considerations: this type MUST NOT be used except as part of a checkgroups control message Published specification: [USEPRO] The content of the "application/news-checkgroups" body part is defined as: checkgroups-body = *( valid-group CRLF ) valid-group = newsgroups-line ; see 5.3 6. Control Messages The following sections document the control messages. "Message" is used herein as a synonym for "article" unless context indicates otherwise. The descriptions below set out REQUIREMENTS to be followed by sites that receive control messages and choose to honour them. However, nothing in these descriptions should be taken as overriding the right of any such site, in accordance with its local policy, to refuse to honour any particular control message, or to refer it to an C. H. Lindsey [Page 14] News Article Architecture and Protocols September 2004 administrator for approval (either as a class or on a case-by-case basis). Relaying Agents MUST propagate all control messages regardless of whether or not they are recognized or processed locally. In the following sections, each type of control message is defined syntactically by defining its verb, its arguments, and possibly its body. 6.1. Digital Signature of Headers [Removed to [USEFOR] (unless it remains in this document as the Primary document, in which case it would not be placed here.] 6.2. Group Control Messages "Group control messages" are the sub-class of control messages that request some update to the configuration of the groups known to a serving agent, namely "newgroup", "rmgroup", "mvgroup" and "checkgroups", plus any others created by extensions to this standard. All of the group control messages MUST have an Approved-header (a-6.14) which, in those hierarchies where appropriate administrative agencies exist (see 1.1), identifies the appropriate person or entity as authorized by those agencies. The authorized person or entity SHOULD adhere to the conventions and restrictions on the format of newsgroup-names established for those hierarchies (see a-5.5). 6.2.1. The 'newgroup' Control Message control-message =/ Newgroup-message Newgroup-message = "newgroup" Newgroup-arguments Newgroup-arguments = CFWS newsgroup-name [ CFWS newgroup-flag ] newgroup-flag = "moderated" The "newgroup" control message requests that the specified group be created or changed. When the request is honoured, if the newgroup- flag "moderated" is present then the status of the group SHOULD be marked as moderated, and vice versa. "Moderated" is the only such flag defined by this standard; other flags MAY be defined for use in cooperating subnets, but newgroup messages containing them MUST NOT be acted on outside of those subnets. NOTE: Specifically, some alternative flags such as "y" and "m", which are sent and recognized by some current software, are NOT part of this standard. Moreover, some existing implementations treat any flag other than "moderated" as indicating an unmoderated newsgroup. Both of these usages are contrary to this standard and control messages with such non-standard flags should be ignored. C. H. Lindsey [Page 15] News Article Architecture and Protocols September 2004 6.2.1.1. The Body of the 'newgroup' Control Message The body of the newgroup message contains the following subparts, preferably in the order shown: 1. An "application/news-groupinfo" part (5.3) containing the name and newsgroups-line of the group. This part MUST be present and SHOULD be used to update any copy of the newsgroups-line (5.3) maintained by the serving agent. 2. Other parts containing useful information about the background of the newgroup message (typically of type "text/plain"). 3. Parts containing initial articles for the newsgroup. See section 6.2.1.2 for details. In the event that there is only the single (i.e. application/news- groupinfo) subpart present, it will suffice to include a "Content- Type: application/news-groupinfo" amongst the headers of the control message. Otherwise, a "Content-Type: multipart/mixed" header will be needed, and each separate part will then need its own Content-Type- header. 6.2.1.2. Initial Articles Some subparts of a "newgroup" or "mvgroup" control message MAY contain an initial set of articles to be posted to the affected newsgroup as soon as it has been created or modified. These parts are identified by having the Media Type "application/news-transmission", possibly with the parameter "usage=inject". The body of each such part should be a complete proto-article, ready for posting. This feature is intended for the posting of charters, initial FAQs and the like to the newly formed group. The Newsgroups-header of the proto-article MUST include the newsgroup-name of the newly created or modified group. It MAY include other newsgroup-names. If the proto-article includes a Message-ID- header, the message identifier in it MUST be different from that of any existing article and from that of the control message as a whole. Alternatively such a message identifier MAY be derived by the injecting agent when the proto-article is posted. The proto-article SHOULD include the header "Distribution: local". The proto-article SHOULD be injected at the serving agent that processes the control message AFTER the newsgroup in question has been created or modified. It MUST NOT be injected if the newsgroup is not, in fact, created (for whatever reason). It MUST NOT be submitted to any relaying agent for transmission beyond the serving agent(s) upon which the newsgroup creation has just been effected (in other words, it is to be treated as having a "Distribution: local" header, whether such a header is actually present or not). C. H. Lindsey [Page 16] News Article Architecture and Protocols September 2004 NOTE: It is not precluded that the proto-article is itself a control message or other type of special article, to be activated only upon creation of the new newsgroup. However, except as might arise from that possibility, any "application/news-transmission" within some nested "multipart/*" structure within the proto-article is not to be activated. 6.2.1.3. Example A "newgroup" with its charter: From: "example.all Administrator" Newsgroups: example.admin.info,example.admin.announce Date: 27 Feb 2002 12:50:22 +0200 Subject: cmsg newgroup example.admin.info moderated Approved: admin@noc.example Control: newgroup example.admin.info moderated Message-ID: MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="nxtprt" Content-Transfer-Encoding: 8bit This is a MIME control message. --nxtprt Content-Type: application/news-groupinfo For your newsgroups file: example.admin.info About the example.* groups (Moderated) --nxtprt Content-Type: application/news-transmission Newsgroups: example.admin.info From: "example.all Administrator" Subject: Charter for example.admin.info Message-ID: Distribution: local Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit The group example.admin.info contains regularly posted information on the example.* hierarchy. --nxtprt-- 6.2.2. The 'rmgroup' Control Message control-message =/ Rmgroup-message Rmgroup-message = "rmgroup" Rmgroup-arguments Rmgroup-arguments = CFWS newsgroup-name The "rmgroup" control message requests that the specified group be removed from the list of valid groups. The Media Type of the body is unspecified; it MAY contain anything, usually an explanatory text. C. H. Lindsey [Page 17] News Article Architecture and Protocols September 2004 NOTE: It is entirely proper for a serving agent to retain the group until all the articles in it have expired, provided that it ceases to accept new articles. 6.2.2.1. Example From: "example.all Administrator" Newsgroups: example.admin.obsolete, example.admin.announce Date: 4 Apr 2002 22:04 -0900 (PST) Subject: cmsg rmgroup example.admin.obsolete Message-ID: Approved: admin@noc.example Control: rmgroup example.admin.obsolete The group example.admin.obsolete is obsolete. Please remove it from your system. 6.2.3. The 'mvgroup' Control Message control-message =/ Mvgroup-message Mvgroup-message = "mvgroup" Mvgroup-arguments Mvgroup-arguments = CFWS newsgroup-name CFWS newsgroup-name [ CFWS newgroup-flag ] The "mvgroup" control message requests that the group specified by the first (old-)newsgroup-name be moved to that specified by the second (new-)newsgroup-name. Thus it is broadly equivalent to a "newgroup" control message for the second group followed by a "rmgroup" control message for the first group. The message body contains an "application/news-groupinfo" part (5.3) containing machine- and human-readable information about the new group, and possibly other subparts as for a "newgroup" control message. The information conveyed in the "application/news-groupinfo" body part, notably its newsgroups-line (5.3), is applied to the new group. When this message is received, the new group is created (if it does not exist already) as for a "newgroup" control message, and SHOULD in any case be made moderated if a newgroup-flag "moderated" is present, and vice versa. At the same time, arrangements SHOULD be made to remove the old group (as with a "rmgroup" control message), but only after a suitable overlap period to allow the network to adjust to the new arrangement. At the same time as a serving agent acts upon this message, all injecting agents associated with that serving agent SHOULD inhibit the posting of new articles to the old group (preferably with some indication to the poster that the new group should have been used). Relaying agents, however, MUST continue to propagate such articles during the overlap period. C. H. Lindsey [Page 18] News Article Architecture and Protocols September 2004 NOTE: It is to be expected that different serving agents will act on this message at different points of time, users of the old group will have to become accustomed to the new arrangement, and followups to already established threads will likely continue under the old group. Therefore, there needs to be an overlap period during which articles may continue to be accepted by relaying and serving agents in either group. This standard does not specify any standard period of overlap (though it would be expected to be expressed in days rather than in months). The inhibition of injection of new articles to the old group may seem draconian, but it is the surest way to prevent the changeover from dragging on indefinitely. Since the "mvgroup" control message is newly introduced in this standard and may not be widely implemented initially, it SHOULD be followed shortly afterwards by a corresponding "newgroup" control message; and again, after a reasonable overlap period, it MUST be followed by a "rmgroup" control message for the old group. In order to facilitate a smooth changeover, serving agents MAY arrange to service requests for access to the old group by providing access to the new group, which would then contain, or appear to contain, all articles posted to either group (including, ideally, the pre-changeover articles from the old one). Nevertheless, if this feature is implemented, the articles themselves, as supplied to reading agents, MUST NOT be altered in any way (and, in particular, their Newsgroups-headers MUST contain exactly those newsgroups present when they were injected). On the other hand, the Xref-header MAY contain entries for either group (or even both). NOTE: Some serving agents that use an "active" file permit an entry of the form "oldgroup xxx yyy =newgroup", which enables any articles arriving for oldgroup to be diverted to newgroup, thus providing a simple implementation of this feature. However, it is known that not all current serving agents will find implementation so easy (especially in the short term) which is why it is not mandated by this standard. Nevertheless, its eventual implementation in all serving agents is to be considered highly desirable. On the other hand, it is recognized that this feature would likely not be implementable if the new group was already in existence with existing articles in it. This situation should not normally arise except when there is already some confusion as to which groups are, or are not, supposed to exist in that hierarchy. Note that the "mvgroup" control message is not really intended to be used for merging two existing groups. 6.2.3.1. Example From: "example.all Administrator" Newsgroups: example.oldgroup,example.newgroup,example.admin.announce Date: 30 Apr 2002 22:04 -0500 (EST) Subject: cmsg mvgroup example.oldgroup example.newgroup moderated C. H. Lindsey [Page 19] News Article Architecture and Protocols September 2004 Message-ID: Approved: admin@noc.example Control: mvgroup example.oldgroup example.newgroup moderated MIME-Version: 1.0 Content-Type: multipart/mixed; boundary=nxt --nxt Content-Type: application/news-groupinfo For your newsgroups file: example.newgroup The new replacement group (Moderated) --nxt The moderated group example.oldgroup is replaced by example.newgroup. Please update your configuration, and please, if possible, arrange to file articles arriving for example.oldgroup as if they were in example.newgroup. --nxt Content-Type: application/news-transmission Newsgroups: example.admin.info From: "example.all Administrator" Subject: Charter for example.newgroup Message-ID: Distribution: local Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit This group (formerly known as example.oldgroup) is for the discussion of examples. --nxt-- 6.2.4. The 'checkgroups' Control Message The "checkgroups" control message contains a list of all the valid groups in a complete hierarchy. control-message =/ Checkgroup-message Checkgroup-message = "checkgroups" Checkgroup-arguments Checkgroup-arguments= [ chkscope ] [ chksernr ] chkscope = 1*( CFWS ["!"] newsgroup-name ) chksernr = CFWS "#" 1*DIGIT A "checkgroups" message applies to any (sub-)hierarchy with a prefix listed in the chkscope parameter, provided that the rightmost matching newsgroup-name in the list is not immediately preceded by a "!". If no chkscope parameter is given, it applies to all hierarchies for which group statements appear in the body of the message. C. H. Lindsey [Page 20] News Article Architecture and Protocols September 2004 NOTE: Some existing software does not support the "chkscope" parameter. Thus a "checkgroups" message SHOULD also contain the groups of other subhierarchies the sender is not responsible for. "New" software MUST ignore groups which do not fall within the chkscope parameter of the "checkgroups" message. The chksernr parameter is a serial number, which can be any positive integer (e.g. just numbered or the date in YYYYMMDD). It SHOULD increase by an arbitrary value with every change to the group list and MUST NOT ever decrease. NOTE: This was added to circumvent security problems in situations where the Date-header cannot be authenticated. Example: Control: checkgroups de !de.alt #248 which includes the whole of the 'de.*' hierarchy, with the exception of its 'de.alt.*' sub-hierarchy. The body of the message has the Media Type "application/news- checkgroups" (5.4). It asserts that the valid-groups it lists are the only newsgroups in the specified hierarchies. NOTE: The checkgroups message is intended to synchronize the list of newsgroups stored by a serving agent, and their newsgroup-descriptions, with the lists stored by other serving agents throughout the network. However, it might be inadvisable for the serving agent actually to create or delete any newsgroups without first obtaining the approval of its administrators for such proposed actions. NOTE: The possibility of removing a complete hierarchy by means of an "invalidation" line beginning with a '!' in the checkgroups-body is no longer provided by this standard. The intent of the feature was widely misunderstood and it was misused more often than it was used correctly. The same effect, if required, can now be obtained by the use of an appropriate chkscope argument in conjunction with an empty checkgroups-body. 6.3. Cancel The cancel message requests that a target article be "canceled", i.e. be withdrawn from circulation or access. control-message =/ Cancel-message Cancel-message = "cancel" Cancel-arguments Cancel-arguments = CFWS msg-id [CFWS] The argument identifies the article to be cancelled by its message identifier. The body SHOULD contain an indication of why the cancellation was requested. The cancel message SHOULD be posted to C. H. Lindsey [Page 21] News Article Architecture and Protocols September 2004 the same newsgroup, with the same distribution, as the article it is attempting to cancel. A serving agent that elects to honour a cancel message SHOULD make the article unavailable for relaying or serving (perhaps by deleting it completely). If the target article is unavailable, and the acceptability of the cancel message cannot be established without it, activation of the cancel message SHOULD be delayed until the target article has been seen. See also sections 7.3 and 7.4. NOTE: It is expected that the security extension envisaged in section a-7.1 will make more detailed provisions for establishing whether honouring a particular cancel message is in order. In particular, it is likely that there will be provision for the digital signature of 3rd party cancels (i.e. those issued other than by the sender, the moderator, or the injector). NOTE: A cancel submitted by the poster for an article in a moderated group will be forwarded to the moderator of that group, and it is up to that moderator to act upon it (7.8). NOTE: The former requirement [RFC 1036] that the From and/or Sender-headers of the cancel message should match those of the original article has been removed from this standard, since it only encouraged cancel issuers to conceal their true identity, and it was not usually checked or enforced by canceling software. Therefore, both the From and/or Sender-headers and any Approved-header should now relate to the entity responsible for issuing the cancel message. 6.4. Ihave, sendme The "ihave" and "sendme" control messages implement a crude batched predecessor of the NNTP [NNTP] protocol. They are largely obsolete on the Internet, but still see use in conjunction with some transport protocols such as UUCP, especially for backup feeds that normally are active only when a primary feed path has failed. There is no requirement for relaying agents that do not support such transport protocols to implement them. NOTE: The ihave and sendme messages defined here have ABSOLUTELY NOTHING TO DO WITH NNTP, despite similarities of terminology. The two messages share the same syntax: control-message =/ Ihave-message Ihave-message = "ihave" Ihave-arguments Ihave-arguments = relayer-name control-message =/ Sendme-message Sendme-message = "sendme" Sendme-arguments Sendme-arguments = Ihave-arguments relayer-name = path-identity ; see a-5.6.1 ihave-body = *( msg-id CRLF ) C. H. Lindsey [Page 22] News Article Architecture and Protocols September 2004 sendme-body = ihave-body The body of the message consists of a list of msg-ids, one per line. [RFC 1036] also permitted the list of msg-ids to appear in the Ihave- or Sendme-arguments with the syntax Ihave-arguments = [FWS] *( msg-id FWS ) [relayer-name] but this form SHOULD NOT now be used, though relaying agents MAY recognize and process it for backward compatibility. The ihave message states that the named relaying agent has received articles with the specified message identifiers, which may be of interest to the relaying agents receiving the ihave message. The sendme message requests that the agent receiving it send the articles having the specified message identifiers to the named relaying agent. Upon receipt of the sendme message, the receiving agent sends the article(s) requested, often (especially when the transport protocol is UUCP) in the form of one or more batches, each containing several articles. The usual form of a batch is defined by the following syntax (which is also used in the application/news transmission media type (5.1)). batch = 1*( batch-header article ) batch-header = "#!" SP rnews SP article-size CRLF rnews = %x72.6E.65.77.73 ; case sensitive "rnews" article-size = 1*DIGIT Thus a batch is a sequence of articles, each prefixed by a header line that includes its size. The article-size is a decimal count of the octets in the article, counting each CRLF as one octet regardless of how it is actually represented. NOTE: Despite the similarity of this format to an executable UNIX script, it is EXTREMELY unwise to feed such a batch into a command interpreter in anticipation of it running a command named "rnews"; the security implications of so doing would be disastrous. These control messages are normally sent essentially as point-to- point messages, by using newsgroup-names in the Newsgroups-header of the form "to." followed by one (or possibly more) components in the form of a relayer-name (see section a-5.5.1 which forbids "to" as the first component of a newsgroup-name). The control message SHOULD then be delivered ONLY to the relaying agent(s) identified by that relayer-name, and any relaying agent receiving such a message which includes its own relayer-name MUST NOT propagate it further. Each pair of relaying agent(s) sending and receiving these messages MUST be immediate neighbors, exchanging news directly with each other. Each relaying agent advertises its new arrivals to the other using ihave messages, and each uses sendme messages to request the articles it lacks. C. H. Lindsey [Page 23] News Article Architecture and Protocols September 2004 To reduce overhead, ihave and sendme messages SHOULD be sent relatively infrequently and SHOULD contain reasonable numbers of message identifiers. If ihave and sendme are being used to implement a backup feed, it may be desirable to insert a delay between reception of an ihave and generation of a sendme, so that a slightly slow primary feed will not cause large numbers of articles to be requested unnecessarily via sendme. 6.5. Obsolete control messages. The following control messages (as described in Appendix A.3) are declared obsolete by this standard: sendsys version whogets senduuname 7. Duties of Various Agents The following section sets out the duties of various agents involved in the creation, relaying and serving of Usenet articles. Insofar as these duties are described as sequences of steps to be followed, it should be understood that it is the effect of these sequences that is important, and implementations may use any method that gives rise to that same effect. In this section, the word "trusted", as applied to the source of some article, means that an agent processing that article has verified, by some means, the identity of that source (which may be another agent or a poster). NOTE: In many implementations, a single agent may perform various combinations of the injecting, relaying and serving functions. Its duties are then the union of the various duties concerned. 7.1. General principles to be followed There are two important principles that news implementors (and administrators) need to keep in mind. The first is the well-known Internet Robustness Principle: Be liberal in what you accept, and conservative in what you send. However, in the case of news there is an even more important principle, derived from a much older code of practice, the Hippocratic Oath (we may thus call this the Hippocratic Principle): First, do no harm. C. H. Lindsey [Page 24] News Article Architecture and Protocols September 2004 It is VITAL to realize that decisions which might be merely suboptimal in a smaller context can become devastating mistakes when amplified by the actions of thousands of hosts within a few minutes. In the case of gateways, the primary corollary to this is: Cause no loops. 7.2. Duties of an Injecting Agent An Injecting Agent is responsible for taking a proto-article from a posting agent and either forwarding it to a moderator or injecting it into the relaying system for access by readers. As such, an injecting agent is considered responsible for ensuring that any article it injects conforms with the rules of [USEFOR]. In the normal course of events, an article that has already been injected into a Netnews network will never pass through another injecting agent. So, if an injecting agent receives an otherwise valid article that has already been injected (as evidenced by the presence of an Injection-Date-header, an Injection-Info-header, or more thath one '%' path-delimiter in a Path-header) it MAY choose to reject it, but otherwise SHOULD cause it to be relayed, as it stands, by a relaying agent (7.3). In exceptional circumstances (e.g. as part of some complex gatewaying process, or where a relaying agent considers it essential for fulfilling its responsibility towards the rest of the network) an already injected article MAY be "reinjected" into the network. This standard does not prescribe any such circumstance; rather this is a matter of policy to be determined by the administrators of each injecting agent, who have the responsibility to ensure that no harm arises. In all other circumstances, unintented reinjection is to be avoided (see 7.9). Nevertheless, in order to preserve the integrity of the network in these special cases, this standard does set out the correct way to reinject. It is usual for an injecting agent to be closely associated with a serving agent, thus giving it access to the list (7.4) showing the moderation status of the newsgroups it is likely to handle. In the event that it does not have such an associated serving agent, it MUST maintain that list itself. 7.2.1. Proto-articles A proto-article SHOULD NOT be propagated in that form to other than injecting agents. A proto-article has the same format as a normal article except that some of the following mandatory headers MAY be omitted: Message-Id- header, Date-header, Path-header (and even From-header if the particular injecting agent can derive that information from other sources). However, if it is intended to offer the proto-article to C. H. Lindsey [Page 25] News Article Architecture and Protocols September 2004 two or more injecting agents in parallel, then it is only the Path- header that MAY be omitted. The headers that can be omitted MUST NOT contain invalid values; they MUST either be correct or not present at all. [Maybe omit that last sentence.] NOTE: An article that is offered for reinjection has, by definition, already been injected once, and is not therefore to be considered as a proto-article. Hence a genuine proto-article will not contain any Injection-Date-header nor any '%' path- delimiter in its Path-header. It MAY contain path-identities with other path-delimiters in the pre-injection portion of its Path-header (a-5.6.3). 7.2.2. Procedure to be followed by Injecting Agents An injecting agent receives proto-articles from posting and followup agents. It verifies them, adds headers where required, and then either forwards them to a moderator or injects them by passing them to serving or relaying agents. It MUST NOT forward an already injected article to a moderator. An injecting agent processes articles as follows: 1. It MUST remove any Injection-Info- or Complaints-To-header already present (though it might be useful to copy them to suitable X- headers). It SHOULD likewise remove any NNTP-Posting-Host, X-Trace or other undocumented tracing header. 2. It SHOULD verify that the article is from a trusted source. However, it MAY allow articles in which headers contain unverified email addresses, that is, addresses which are not known to be valid for the trusted source, and notably so if they end in ".invalid". 3. It SHOULD reject any article whose Date-header (a-5.1) is more than 24 hours into the future (and MAY use a margin less than that 24 hours). It MUST, except when reinjecting, reject any article with an Injection-Date-header already present (and SHOULD do likewise with any NNTP-Posting-Date-header). It MAY when reinjecting, but only if there is no Injection-Date-header present, reject any article whose Date-header appears to be stale (e.g. more than 72 hours into the past). 4. It MUST reject any article that does not have the correct mandatory headers for a proto-article (or, when reinjecting, all the mandatory headers other than Injection-Date), or which contains any header that does not have legal contents. It SHOULD reject any article which contains any header deprecated for Netnews (a-4.2.1). It SHOULD reject any article whose Newsgroups-header does not contain at least one newsgroup-name for an existing group (as listed by its associated serving agent) and it MAY reject any newsgroup-name which, although syntactically correct, violates a policy restriction established, for some C. H. Lindsey [Page 26] News Article Architecture and Protocols September 2004 (sub-)hierarchy, by an agency with the appropriate authority (1.2). Observe that crossposting to unknown newsgroups is not precluded provided at least one of those in the Newsgroups-header is listed. NOTE: This ability to reject newsgroup-names in breach of established policy does not extend to relaying agents, though it might be reasonable for posting agents to do it. 5. If the article is rejected (for reasons given above, or for other formatting errors or matters of site policy) the posting agent SHOULD be informed (such as via an NNTP 44x response code) that posting has failed and the article MUST NOT then be processed further. 6. The Message-ID, Date and From-headers (and their contents) MUST be added when not already present (but that situation could not arise during reinjection). A User-Agent-header MAY be added (or an already present User-Agent-header MAY be augmented) so as to identify the software (e.g. "INN/1.7.2") used by the injecting agent. 7. The injecting agent MUST NOT alter the body of the article in any way (including any change of Content-Transfer-Encoding). It MAY, except when reinjecting, add other headers not already provided by the poster, but SHOULD NOT alter, delete, or reorder any existing header, with the specific exception of "tracing" headers such as Injection-Info and Complaints-To, which are to be removed as already mentioned. It MAY also, as an interim measure pending widespread adoption of the newly introduced (a-5.5) folding whitespace, reformat the Newsgroups- and any Followup-To-header by removing any such whitespace inserted by the posting agent. 8. If the Newsgroups-header contains one or more moderated groups and the article does NOT contain an Approved-header, the injecting agent MUST forward it to a moderator as specified in section 7.2.3 below. 9. Otherwise, a Path-header with a tail-entry (a-5.6.3) MUST be correctly added if not already present. During reinjection, the existing Path-header SHOULD be retained. 10.It MUST then prepend the path-identity of the injecting agent and a '%' path-delimiter (which serves to separate the pre-injection and post-injection regions of the Path-content) to the Path- content; this SHOULD then be followed by CRLF and WSP if it would otherwise result in a line longer than 79 characters. The prepended path-identity MUST be an FQDN mailable address (a- 5.6.2). This could result in more that one '%' path-delimiter in the case of reinjection. See a-5.6.4 for the significance of the various path-delimiters. C. H. Lindsey [Page 27] News Article Architecture and Protocols September 2004 11.An Injection-Info-header (a-6.19) SHOULD be added, identifying the trusted source of the article, and a suitable Complaints-To-header (a-6.20) MAY be added. Each injecting agent SHOULD use a consistent form of the Injection-Info-header for all articles emanating from the same or similar origins. NOTE: The step above is the only place in which an Injection- Info- or Complaints-To-header is to be created. It follows that these headers MUST NOT be created, replaced, changed or deleted by any other agent (except during reinjection, in which case they will always relate to the latest injection and can, to that extent, be regarded as variant headers). 12.The injecting agent MUST then add an Injection-Date-header (a-5.7) if one is not already present, but it MUST NOT alter, or delete, an already present Injection-Date-header (and likewise SHOULD NOT alter, or delete, an already present NNTP-Posting-Date-header). Finally, it forwards the article to one or more relaying or serving agents, and the injection process is to be considered complete. NOTE: The step above is the only place where an Injection-Date- header is to be created It follows that it MUST NOT subsequently be replaced, changed or deleted by any other agent, even during reinjection. 7.2.3. Procedure for Forwarding to a Moderator An injecting agent forwards ar article to a moderator as follows: 1. It MUST forward it to the moderator of the first (leftmost) moderated group listed in the Newsgroups-header via email (see 7.8 for how that moderator may forward it to further moderators). There are two possibilities for doing this: (a) The complete article is encapsulated (headers and all) within the email, preferably using the Content-Type "application/news-transmission" (5.1) with any usage parameter set to "moderate". Moreover, there SHOULD NOT be more than one encapsulated article within the one email. This method has the advantage of removing any possible conflict between Netnews and Email headers, or of changes to those headers during transport through email. (b) The article is sent as an email as it stands, with the addition of such extra headers (e.g. a To-header) as are necessary for an email. The existing Message-ID-header SHOULD be retained. Although both of these methods have seen use in the past, the preponderance of current usage on Usenet has been for method (b) and many moderators are ill-prepared to deal with method (a). Therefore, method (a) SHOULD NOT be used until such time as the majority of moderators are able to accept it. C. H. Lindsey [Page 28] News Article Architecture and Protocols September 2004 2. This standard does not prescribe how the email address of the moderator is to be determined, that being a matter of policy to be arranged by the agency responsible for the oversight of each hierarchy. Nevertheless, there do exist various agents worldwide which provide the service of forwarding to moderators, and the address to use with them is obtained as follows: (a) Each '.' in the newsgroup-name is replaced with a '-'. (b) The result of these operations is used as the local-part of the mailbox of the agent. For example, articles intended for "news.announce.important" would be emailed to "news- announce-important@forwardingagent.example". 7.3. Duties of a Relaying Agent A Relaying Agent accepts injected articles from injecting and other relaying agents and passes them on to relaying or serving agents according to mutually agreed policy. Relaying agents SHOULD accept articles ONLY from trusted agents. An article SHOULD NOT be relayed unless the sending agent has been configured to supply and the receiving agent to receive at least one of the newsgroup-names in its Newsgroups-header and at least one of the distributions in its Distribution-header, if any. Exceptionally, ALL relaying agents are deemed willing to supply or accept the distribution "world", and NO relaying agent should supply or accept the distribution "local". [That SHOULD has been demoted from a MUST in draft-13. Any objections?] NOTE: Although it would seem redundant to filter out unwanted distributions at both ends of a relaying link (and it is clearly more efficient to do so at the sending end), many sending sites have been reluctant, historically speaking, to apply such filters (except to ensure that distributions local to their own site or cooperating subnet did not escape); moreover they tended to configure their filters on an "all but those listed" basis, so that new and hitherto unheard of distributions would not be caught. Indeed many "hub" sites actually wanted to receive all possible distributions so that they could feed on to their clients in all possible geographical (or organizational) regions. Therefore, it is desirable to provide facilities for rejecting unwanted distributions at the receiving end. Indeed, it may be simpler to do so locally than to inform each sending site of what is required, especially in the case of specialized distributions (for example for control messages, such as cancels from certain issuers) which might need to be added at short notice. The possibility for reading agents to filter distributions has been provided (7.7) for the same reason. C. H. Lindsey [Page 29] News Article Architecture and Protocols September 2004 An article SHOULD NOT be relayed if the path-identity of the receiving agent (or some known alias thereof) appears in its Path- header, and the receiving agent MAY detect whether its own path- identity is already present in the Path-content so as to avoid further unnecessary relaying. [See related remarks under serving agents.] A relaying agent processes articles as follows: 1. It MUST establish the trusted identity of the source of the article and compare it with the leftmost path-identity of the Path-content. If it matches it MUST then prepend its own path- identity and a '/' path-delimiter to the Path-content; this SHOULD then be followed by CRLF and WSP if it would otherwise result in a line longer than 79 characters. If it does not match then it prepends instead two entries to the Path-content; firstly the true established path-identity of the source followed by a '?' path- delimiter, and then, to the left of that, its own path-identity followed by a '/' path-delimiter as usual. This prepending of two entries SHOULD NOT be done if the provided and established identities match. See a-5.6.4 for the significance of the various path-delimiters. NOTE: In order to prevent overloading, relaying agents should not routinely query an external entity (such as a DNS-server) in order to verify an article (though a local cache of the required information might usefully be consulted). 2. It MUST examine the Injection-Date-header (or, if that is absent, the Date-header) and reject the article as stale (a-5.7) if that predates the earliest articles of which it normally keeps record, or if it is more than 24 hours into the future (the margin MAY be less than that 24 hours). 3. It MUST reject any article that does not have the correct mandatory headers (section a-5) present with legal contents. 4. It SHOULD reject any article whose optional headers (section a-6) do not have legal contents. [Is that too strong? Are relaying agents really expected to check headers in that detail? I suggest s/SHOULD/MAY/. Even the MUST in Step 4 for mandatory headers might be demoted to SHOULD.] 5. It SHOULD reject any article that has already been sent to it (a database of message identifiers of recent messages is usually kept and matched against). 6. It SHOULD reject any article that matches an already received cancel message (or an equivalent Supersedes-header) issued by its poster or by some other trusted entity. 7. It MAY reject any article without an Approved-header posted to newsgroups known to be moderated (this practice is strongly recommended, but the information necessary to do so may not be C. H. Lindsey [Page 30] News Article Architecture and Protocols September 2004 available to all agents). 8. It MAY delete any Xref-header that is present. 9. Finally, it passes the articles on to neighbouring relaying and serving agents. If the article is rejected as being invalid, unwanted or unacceptable due to site policy, the agent that passed the article to the relaying agent SHOULD be informed (such as via an NNTP 43x response code) that relaying failed. In order to prevent a large number of error messages being sent to one location, relaying agents MUST NOT inform any other external entity that an article was not relayed UNLESS that external entity has explicitly requested that it be informed of such errors. Relaying agents MUST NOT alter, delete or rearrange any part of an article expect for headers designated as variant (a-4.2.5.3). In particular o they MUST NOT create or augment a User-Agent-header in order to identify themselves; o they MUST NOT rewrite the Newsgroups-header in any way, even if some supposedly non-existent newsgroup is included; o they MUST NOT refold any header (i.e. they must pass on the folding as received), even to remove FWS from a Newsgroups- header; o they MUST NOT alter the Date-header or the Injection-Date-header; o they MUST NOT delete any unrecognized header whose header-name is syntactically correct (whether or not it is registered with IANA [RFC 3864]); o they MUST NOT change the Content-Transfer-Encoding of the body or any body part. 7.3.1. Path-Header Example Path: foo.isp.example/ foo-server/bar.isp.example?10.123.12.2/old.site.example! barbaz/baz.isp.example%dialup123.baz.isp.example!not-for-mail NOTE: That article was injected into the news stream by baz.isp.example (complaints may be addressed to abuse@baz.isp.example). The injector has taken care to record that it got it from dialup123.baz.isp.example. "not-for-mail" is a dummy tail-entry, though sometimes a real userid is put there. The article was relayed, perhaps by UUCP, to the machine known, at least to old.site.example, as "barbaz". Barbaz relayed it to old.site.example, which does not yet conform to this standard (hence the '!' path-delimiter). So one cannot be sure that it really came from barbaz. C. H. Lindsey [Page 31] News Article Architecture and Protocols September 2004 Old.site.example relayed it to a site claiming to have the IP address [10.123.12.2], and claiming (by using the '/' path- delimiter) to have verified that it came from old.site.example. [10.123.12.2] relayed it to "foo-server" which, not being convinced that it truly came from [10.123.12.2], did a reverse lookup on the actual source and concluded it was known as bar.isp.example (that is not to say that [10.123.12.2] was not a correct IP address for bar.isp.example, but simply that that connection could not be substantiated by foo-server). Observe that foo-server has now added two entries to the Path. "foo-server" is a locally significant name within the complex site of many machines run by foo.isp.example, so the latter should have no problem recognizing foo-server and using a '/' path-delimiter. Presumably foo.isp.example then delivered the article to its direct clients. It appears that foo.isp.example and old.site.example decided to fold the line, on the grounds that it seemed to be getting a little too long. 7.4. Duties of a Serving Agent A Serving Agent takes an article from a relaying or injecting agent and files it in a "news database". It also provides an interface for reading agents to access the news database. This database is normally indexed by newsgroup with articles in each newsgroup identified by an article-locator (usually in the form of a decimal number - see a- 6.16). A serving agent MUST maintain a list of the newsgroups it stores in its news database showing the moderation status of each one (see 6.2.1), and SHOULD include in that list all groups likely to be crossposted to from those groups (e.g. all other groups in the same hierarchy(ies)). NOTE: Since control messages are often of interest, but should not be displayed as normal articles in regular newsgroups, it is common for serving agents to make them available in a pseudo- newsgroup named "control" or in a pseudo-newsgroup in a sub- hierarchy under "control." (e.g. "control.cancel"). A serving agent MAY decline to accept an article if its own path- identity is already present in the Path-content or if the Path- content contains some path-identity whose articles the serving agent does not want, as a matter of local policy. [That has been changed from a-5.6.1 in previous drafts, where it said "A relaying agent MAY decline...". That seemed plain wrong to me. It is fine for a relaying agent to ignore articles which have apparently already passed through it (and I still say that), but surely it is for serving agents to "reject" for policy reasons, and to which the NOTE below would apply. Comments?] C. H. Lindsey [Page 32] News Article Architecture and Protocols September 2004 NOTE: This last facility is sometimes used to detect and decline control messages (notably cancel messages) which have been deliberately seeded with a path-identity to be "aliased out" by sites not wishing to act upon them. [Again, is this "aliasing out" usually detected by the serving agent, or does it more usually work because the previous relaying agent will never have sent it in the first place?] A serving agent processes articles as follows: 1. It MUST establish the trusted identity of the source of the article and modify the Path-header as for a relaying agent (7.3). 2. It MUST examine the Injection-Date-header (or, if that is absent, the Date-header) and reject the article as stale (a-5.7) if that predates the earliest articles of which it normally keeps record, or if it is more than 24 hours into the future (the margin MAY be less than that 24 hours). 3. It MUST reject any article that does not have the correct mandatory headers (section a-5) present, or which contains any header that does not have legal contents. 4. It SHOULD reject any article that has already been sent to it (a database of message identifiers of recent articles is usually kept and matched against). 5. It SHOULD reject any article that matches an already received cancel message (or an equivalent Supersedes-header) issued by its poster or by some other trusted entity. 6. It MUST reject any article without an Approved-header posted to any newsgroup listed as moderated. 7. It MUST remove any Xref-header (a-6.16) from each article. It then MAY (and usually will) generate a fresh Xref-header. 8. Finally, it stores the article in its news database. Serving agents MUST NOT create new newsgroups simply because an unrecognized newsgroup-name occurs in a Newsgroups-header (see a- 7.2.1 for the correct method of newsgroup creation). Serving agents MUST NOT alter, delete or rearrange any part of an article in any other way. The list of particular cases given for relaying agents (7.3) applies here also. 7.5. Duties of a Posting Agent A Posting Agent is used to assist the poster in creating a valid proto-article and forwarding it to an injecting agent. C. H. Lindsey [Page 33] News Article Architecture and Protocols September 2004 Postings agents SHOULD ensure that proto-articles they create are valid according to [USEFOR] and other applicable policies. In particular, they MUST NOT create any Injection-Date-, Injection-Info- or Complaints-To-header. Posting agents meant for use by ordinary posters SHOULD reject any attempt to post an article which cancels or Supersedes another article of which the poster is not the author. 7.6. Duties of a Followup Agent A Followup Agent is a special case of a posting agent, and as such is bound by all the posting agent's requirements. Followup agents MUST create valid followups and are subject to special requirements involving certain inheritable (a-4.2.5.2) and other headers. Wherever in the following it is stated that, "by default", a header is to be taken from some header in the precursor, it means that its initial content (plus its extension-parameters, if any) are to be copied from those of that precursor header. However, posters MAY then override that default before posting if they so wish. NOTE: There is no provision in this standard for a followup to have more than one precursor (though it might be permitted in some future extension). 1. The Newsgroups-header (a-5.5) SHOULD by default be taken from the precursor's Followup-To-header if present, and otherwise from the precursor's Newsgroups-header. However, if the content of that Followup-To-header consists of "poster" (and the user does not override it), then the followup MUST NOT be posted but, rather, is to be emailed to the precursor's poster. 2. The Subject-header SHOULD by default be taken from that of the precursor. The case sensitive string "Re: " MAY be prepended to its Subject-Content unless it already begins with that string. 3. The Distribution-header (a-6.6) SHOULD by default be taken from the precursor's Distribution-header, if any. 4. If the precursor did not have a References-header (a-6.10), the followup's References-content MUST be formed by the message identifier of the precursor. A followup to an article which had a References-header MUST have a References-header containing the precursor's References-content (subject to trimming as described below) plus the precursor's message identifier appended to the end of the list (separated from it by CFWS). Followup agents MAY trim References-headers which have grown to excessive length, but the first and last message identifiers from the precursor MUST NOT be removed. 5. If the precursor contains a Mail-Copies-To-header (a-6.8), the actions to be taken, in accordance with the Mail-Copies-To- content, (and subject to manual override by the poster) are as C. H. Lindsey [Page 34] News Article Architecture and Protocols September 2004 follows: "nobody" (or when the header is absent) The followup agent SHOULD NOT email a copy of a posted followup to the poster of the precursor. "poster" The followup agent SHOULD (if it has the necessary capability) email a copy of a posted followup, which MUST then be sent to the address(es) in the Reply-To-header, and in the absence of that to the address(es) in the From- header. a copy-addr The followup agent SHOULD likewise email a copy of a posted followup, which MUST then be sent to the copy-addr. When emailing a copy, the followup agent SHOULD also include a "Posted-And-Mailed: yes" header (a-6.9). Followup agents SHOULD NOT attempt to send email to any address ending in ".invalid". 7.7. Duties of a Reading Agent A reading agent downloads articles from a serving agent, as directed by the reader, and displays them (or processes them in some other manner). The article as displayed MUST be identical to the article as originally posted, subject only to limitations of the display device (such as availability of charsets, etc.). It MUST provide facilities for decoding any Content-Transfer-Encodings, encoded- words, etc., but SHOULD also have the capability to show the article exactly as received. It MAY present lists of articles available for display, and MAY structure those lists so as to show the relationships between the articles, as determined by the References-, Subject-, Date- and other-headers (see [USEAGE] for some usual methods of doing this). It MAY also be configured so that unwanted distributions (a-6.6) are ignored. 7.8. Duties of a Moderator A Moderator receives news articles by email, decides whether to accept them and, if so, either injects them into the news stream or forwards them to further moderators. Articles will be received by the moderator either encapsulated as an object of Content-Type application/news-transmission (or possibly encapsulated but without an explicit Content-Type-header), or else directly as an email already containing all the headers appropriate for a Netnews article (see 7.2.2). Moderators SHOULD be prepared to accept articles in either format. C. H. Lindsey [Page 35] News Article Architecture and Protocols September 2004 A moderator processes an article, as submitted to any newsgroup that he moderates, as follows: 1. He decides, on the basis of whatever moderation policy applies to his group, whether to accept or reject the article. He MAY do this manually, or else partially or wholly with the aid of appropriate software for whose operation he is then responsible. If the article is a cancel nessage (6.3) issued by the poster of an earlier article, then he is expected to cancel that earlier article (in which case there is no more to be done). He MAY modify the article if that is in accordance with the applicable moderation policy (and in particular he MAY remove redundant headers and add Comments and other informational headers). He also needs to be aware if any change he makes to the article will invalidate some authentication check provided by the poster or by an earlier moderator. If the article is rejected, then it normally fails for all the newsgroups for which it was intended. If it is accepted, the moderator proceeds with the following steps. 2. If the Newsgroups-header contains further moderated newsgroups for which approval has not already been given, he adds an indication (identifying both himself and the name of the group) that he approves the article, and then forwards it to the moderator of the leftmost unapproved group (which, if this standard has been followed correctly, will generally be the next moderated group to the right of his own). There are two ways to do this: (a) He emails it to the submission address of the next moderator (see section 7.2.2 for the proper method of doing this), or (b) he rotates the newsgroup-names in the Newsgroups-header to the left so that the targeted group is the leftmost moderated group in that header, and injects it as below (thus causing the injecting agent to email it to the correct moderator). However, he MUST first ensure that the article contains no Approved-header. NOTE: This standard does not prescribe how a moderator's approval is to be indicated (though a future standard may do so). Possible methods include adding an Approved header (or a similar but differently named header if method (b) is being used) listing all the approvals made so far, or adding a separate header for each individual approval (the header X-Auth is sometimes used for this purpose). The approval may also be confirmed with some form of digital signature (a-7.1). 3. If the Newsgroups-header contains no further unapproved moderated groups, he adds an Approved-header (a-6.14) identifying himself and, insofar as is possible, all the other moderators who have approved the article. He thus assumes responsibility for having ensured that the article was acceptable to the moderators of all the moderated groups involved. C. H. Lindsey [Page 36] News Article Architecture and Protocols September 2004 4. The Date-header SHOULD be retained. Any Injection-Date-header already present (though there should be none) MUST be removed. Exceptionally, if it is known that the injecting agent does not yet support the Injection-Date-header and the Date-header appears to be stale (a-5.7) for reasons understood by the moderator (e.g. delays in the moderation process) he MAY substitute the current date. The Message-ID-header SHOULD also be retained unless it is obviously non-compliant with this standard. NOTE: A message identifier created by a conforming posting or injecting agent, or even by a mail user agent conforming to [RFC 2822], may reasonably be supposed to be conformant (and will, in any case, be caught by the injecting agent if it is not). 5. Any variant headers (a-4.2.5.3) MUST be removed, except that a Path-header MAY be truncated to only its pre-injection region (a- 5.6.3). Any Injection-Info-header (a-6.19) or Complaints-To- header (a-6.20) SHOULD be removed (and if they are not, the injecting agent will do so, as required in 7.2.2). 6. He then causes the article to be injected, having first observed all the duties of a posting agent. NOTE: This standard does not prescribe how the moderator or moderation policy for each newsgroup is established; rather it assumes that whatever agencies are responsible for the relevant network or hierarchy (1.1) will have made appropriate arrangements in that regard. 7.9. Duties of a Gateway A Gateway transforms an article into the native message format of another medium, or translates the messages of another medium into news articles. Encapsulation of a news article into a message of MIME type application/news-transmission, or the subsequent undoing of that encapsulation, is not gatewaying, since it involves no transformation of the article. There are two basic types of gateway, the Outgoing Gateway that transforms a news article into a different type of message, and the Incoming Gateway that transforms a message from another medium into a news article and injects it into a news system. These are handled separately below. The primary dictat for a gateway is: Above all, prevent loops. Transformation of an article into another medium stands a very high chance of discarding or interfering with the protection inherent in the news system against duplicate articles. The most common problem caused by gateways is "spews," gateway loops that cause previously posted articles to be reinjected repeatedly into Usenet. To prevent this, a gateway MUST take precautions against loops, as detailed C. H. Lindsey [Page 37] News Article Architecture and Protocols September 2004 below. If bidirectional gatewaying (both an incoming and an outgoing gateway) is being set up between Netnews and some other medium, the incoming and outgoing gateways SHOULD be coordinated to avoid unintended reinjection of gated articles. Circular gatewaying (gatewaying a message into another medium and then back into Netnews) SHOULD NOT be done; encapsulation of the article SHOULD be used instead where this is necessary. A second general principal of gatewaying is that the transformations applied to the message SHOULD be as minimal as possible while still accomplishing the gatewaying. Every change made by a gateway potentially breaks a property of one of the media or loses information, and therefore only those transformations made necessary by the differences between the media should be applied. It is worth noting that safe bidirectional gatewaying between a mailing list and a newsgroup is far easier if the newsgroup is moderated. Posts to the moderated group and submissions to the mailing list can then go through a single point that does the necessary gatewaying and then sends the message out to both the newsgroup and the mailing list at the same time, eliminating most of the possibility of loops. Bidirectional gatewaying between a mailing list and an unmoderated newsgroup, in contrast, is difficult to do correctly and is far more fragile. Newsgroups intended to be bidirectionally gated to a mailing list SHOULD therefore be moderated where possible, even if the moderator is a simple gateway and injecting agent that correctly handles crossposting to other moderated groups and otherwise passes all traffic. 7.9.1. Duties of an Outgoing Gateway From the perspective of Netnews, an outgoing gateway is just a special type of reading agent. The exact nature of what the outgoing gateway will need to do to articles depends on the medium to which the articles are being gated. The operation of the outgoing gateway is only subject to additional constraints in the presence of one or more corresponding incoming gateways back from that medium to Netnews, since this opens the possibility of loops. In general, the following practices are recommended for all outgoing gateways, regardless of whether there is known to be a related incoming gateway, both as a precautionary measure and as a guideline to quality of implementation. 1. The message identifier of the news article should be preserved if at all possible, preferably as or within the corresponding unique identifier of the other medium, but if not at least as a comment in the message. This helps greatly with preventing loops. C. H. Lindsey [Page 38] News Article Architecture and Protocols September 2004 2. The Date and Injection-Date of the news article should also be preserved if possible, for similar reasons. 3. The message should be tagged in some way so as to prevent its reinjection into Netnews. This may be impossible to do without knowledge of potential incoming gateways, but it is better to try to provide some indication even if not successful; at the least, a human-readable indication that the article should not be gated back to Netnews can help locate a human problem. 4. Netnews control messages should not be gated to another medium unless they would somehow be meaningful in that medium. 5. Changes MAY be made to the Content-Transfer-Encoding of some or all parts of the body, and even to the charsets specified in encoded-words or in Content-Type-headers, but such changes SHOULD NOT be made unless absolutely necessary. 7.9.2. Duties of an Incoming Gateway The incoming gateway has the serious responsibility of ensuring that all of the requirements of this standard are met by the articles that it forms. In addition to its special duties as a gateway, it bears all of the duties and responsibilities of an injecting agent as well, and additionally has the same responsibility of a relaying agent to reject articles that it has already gatewayed. An incoming gateway MUST NOT gate the same message twice. It may not be possible to ensure this in the face of mangling or modification of the message, but at the very least a gateway, when given a copy of a message it has already gated identical except for trace headers (like Received in Email or Path in Netnews) MUST NOT gate the message again. An incoming gateway SHOULD take precautions against having this rule bypassed by modifications of the message that can be anticipated. News articles prepared by gateways MUST be legal news articles. In particular, they MUST include all of the mandatory headers, MUST fully conform to the restrictions on said headers, and SHOULD exclude any deprecated headers (a-4.2.1). This often requires that a gateway function not only as a relaying agent, but also partly as a posting agent, aiding in the synthesis of a conforming article from non- conforming input. Incoming gateways MUST NOT pass control messages (articles containing a Control- or Supersedes-header) without removing or renaming that header. Gateways MAY, however, generate their own cancel messages, under the general allowance for injecting agents to cancel their own messages ([USEAGE]). If a gateway receives a message that it can determine is a valid equivalent of a cancel message in the medium it is gatewaying, it SHOULD discard that message without gatewaying it, generate a corresponding cancel message of its own, and inject that cancel message. C. H. Lindsey [Page 39] News Article Architecture and Protocols September 2004 Incoming gateways MUST NOT inject control messages other than cancels. Encapsulation SHOULD be used instead of gatewaying, when direct posting is not possible or desirable. NOTE: It is not unheard of for mail-to-news gateways to be used to post control messages, but encapsulation should be used for these cases instead. Gateways by their very nature are particularly prone to loops. Spews of normal articles are bad enough; spews of control messages with special significance to the news system, possibly resulting in high processing load or even email sent for every message received, are catastrophic. It is far preferable to construct a system specifically for posting control messages that can do appropriate consistency checks and authentication of the originator of the message. If there is a message identifier that fills a role similar to that of the Message-ID-header in news, it SHOULD be used in the formation of the message identifier of the news article, perhaps with transformations required to meet the uniqueness requirement of Netnews and with the removal of any comments so as to comply with the syntax in section a-5.3. Such transformations SHOULD be designed so that two messages with the same identifier generate the same Message-ID-header. NOTE: Message identifiers play a central role in the prevention of duplicates, and their correct use by gateways will do much to prevent loops. Netnews does, however, require that message identifiers be unique, and therefore message identifiers from other media may not be suitable for use without modification. A balance must be struck by the gateway between preserving information used to prevent loops and generating unique message identifiers. Exceptionally, if there are multiple incoming gateways for a particular set of messages, each to a different newsgroup(s), each one SHOULD generate a message identifier unique to that gateway. Each incoming gateway nonetheless MUST ensure that it does not gate the same message twice. NOTE: Consider the example of two gateways of a given mailing list into the world-wide Usenet newsgroups, both of which preserve the email message identifier. Each newsgroup may then receive a portion of the messages (different sites seeing different portions). In these cases, where there is no one "official" gateway, some other method of generating message identifiers has to be used to avoid collisions. It would obviously be preferable for there to be only one gateway which crossposts, but this may not be possible to coordinate. If no date information is available, the gateway MAY supply a Date- header with the gateway's current date. If no injection-date information is available, the gateway MUST supply an Injection-Date- header with whatever date information is available, and otherwise with the gateway's current date. If only partial information is C. H. Lindsey [Page 40] News Article Architecture and Protocols September 2004 available (e.g. date but not time), this SHOULD be fleshed out to a full Date- and/or Injection-Date-header by adding default values rather than discarding this information. Only in very exceptional circumstances should Date information be discarded, as it plays an important role in preventing reinjection of old messages. An incoming gateway MUST add a Sender-header to the news article it forms containing the mailbox of the administrator of the gateway. Problems with the gateway may be reported to this mailbox. The display-name portion of this mailbox SHOULD indicate that the entity responsible for injection of the message is a gateway. If the original message already had a Sender-header, it SHOULD be renamed so that its contents can be preserved. 7.9.3. Example To illustrate the type of precautions that should be taken against loops, here is an example of the measures taken by one particular combination of mail-to-news and news-to-mail gateways at Stanford University designed to handle bidirectional gatewaying between mailing lists and unmoderated groups. 1. The news-to-mail gateway preserves the message identifier of the news article in the generated email message. The mail-to-news gateway likewise preserves the email message identifier provided that it is syntactically valid for Netnews. This allows the news system's built-in suppression of duplicates to serve as the first line of defense against loops. 2. The news-to-mail gateway adds an X-Gateway-header to all messages it generates. The mail-to-news gateway discards any incoming messages containing this header. This is robust against mailing list managers that replace the message identifier, and against any number of email hops, provided that the other message headers are preserved. 3. The mail-to-news gateway inserts the host name from which it received the email message in the pre-injection region of the Path (a-5.6.3). The news-to-mail gateway refuses to gateway any message that contains the list server name in the pre-injection region of its Path-header. This is robust against any amount of munging of the message headers by the mailing list, provided that the email only goes through one hop. 4. The mail-to-news gateway is designed never to generate bounces to the envelope sender. Instead, articles that are rejected by the news server (for reasons not warranting silent discarding of the message) result in a bounce message sent to an errors address known not to forward to any mailing lists, so that they can be handled by the news administrators. These precautions have proven effective in practice at preventing loops for this particular application (bidirectional gatewaying between mailing lists and locally distributed newsgroups where both C. H. Lindsey [Page 41] News Article Architecture and Protocols September 2004 gateways can be designed together). General gatewaying to world-wide newsgroups poses additional difficulties; one must be very wary of strange configurations, such as a newsgroup gated to a mailing list which is in turn gated to a different newsgroup. 8. Security and Related Considerations There is no security. Don't fool yourself. Usenet is a prime example of an Internet Adhocratic-Anarchy; that is, an environment in which trust forms the basis of all agreements. It works. 8.1. Leakage Articles which are intended to have restricted distribution are dependent on the goodwill of every site receiving them. The "Archive: no" header (a-6.12) is available as a signal to automated archivers not to file an article, but that cannot be guaranteed. The Distribution-header makes provision for articles which should not be propagated beyond a cooperating subnet. The key security word here is "cooperating". When a machine is not configured properly, it may become uncooperative and tend to distribute all articles. The flooding algorithm is extremely good at finding any path by which articles can leave a subnet with supposedly restrictive boundaries, and substantial administrative effort is required to avoid this. Organizations wishing to control such leakage are strongly advised to designate a small number of official gateways to handle all news exchange with the outside world (however, making such gateways too restrictive can also encourage the setting up of unofficial paths which can be exceedingly hard to track down). The sendme control message (6.4), insofar as it is still used, can be used to request articles with a given message identifier, even one that is not supposed to be supplied to the requestor. 8.2. Attacks 8.2.1. Denial of Service The proper functioning of individual newsgroups can be disrupted by the massive posting of "noise" articles, by the repeated posting of identical or near identical articles, by posting followups unrelated to their precursors, or which quote their precursors in full with the addition of minimal extra material (especially if this process is iterated), and by crossposting to, or setting followups to, totally unrelated newsgroups. Many have argued that "spam", massively multiposted (and to a lesser extent massively crossposted) articles, usually for advertising purposes, also constitutes a DoS attack in its own regard. This may be so. C. H. Lindsey [Page 42] News Article Architecture and Protocols September 2004 Such articles intended to deny service, or other articles of an inflammatory nature, may also have their From or Reply-To addresses set to valid but incorrect email addresses, thus causing large volumes of email to descend on the true owners of those addresses. Similar effects could be caused by any email header which could cause every reading agent receiving it to take some externally visible action. For example, the Disposition-Notification-To-header defined in [RFC 2298] could cause huge numbers of acknowledgements to be emailed to an unsuspecting third party (for which reason [RFC 2298] declares that that header SHOULD NOT be used in Netnews). It is a violation of this standard for a poster to use as his address a mailbox which he is not entitled to use. Even addresses with an invalid local-part but a valid domain can cause disruption to the administrators of such domains. Posters who wish to remain anonymous or to prevent automated harvesting of their addresses, but who do not care to take the additional precautions of using more sophisticated anonymity measures, should avoid that violation by the use of addresses ending in the ".invalid" top-level-domain (see a-5.2). A malicious poster may also prevent his article being seen at a particular site by preloading that site into the Path-header (a- 5.6.1) and may thus prevent the true owner of a forged From or Reply-To address from ever seeing it. A malicious complainer may submit a modified copy of an article (e.g. with an altered Injection-Info-header) to the administrator of an injecting agent in an attempt to discredit the author of that article and even to have his posting privileges removed. Administrators should therefore obtain a genuine copy of the article from their own serving agent before taking such precipitate action. Administrative agencies with responsibility for establishing policies in particular hierarchies can and should set bounds upon the behaviour that is considered acceptable within those hierarchies (for example by promulgating charters for individual newsgroups, and other codes of conduct). Whilst this standard places an onus upon injecting agents to bear responsibility for the misdemeanours of their posters (which includes non-adherence to established policies of the relevant hierarchies as provided in section 7.2), and to provide assistance to the rest of the network by making proper use of the Injection-Info- (a-6.19) and Complaints-To- (a-6.20) headers, it makes no provision for enforcement, which may in consequence be patchy. Nevertheless, injecting sites which persistently fail to honour their responsibilities or to comply with generally accepted standards of behaviour are likely to find themselves blacklisted, with their articles refused propagation and even subject to cancellation, and other relaying sites would be well advised to withdraw peering arrangements from them. C. H. Lindsey [Page 43] News Article Architecture and Protocols September 2004 8.2.2. Compromise of System Integrity The posting of unauthorized (as determined by the policies of the relevant hierarchy) control messages can cause unwanted newsgroups to be created, or wanted ones removed, from serving agents. Administrators of such agents SHOULD therefore take steps to verify the authenticity of such control messages, either by manual inspection (particularly of the Approved-header) or by checking any digital signatures that may be provided (see a-7.1). In addition, they SHOULD periodically compare the newsgroups carried against any regularly issued checkgroups messages, or against lists maintained by trusted servers and accessed by out-of-band protocols such as FTP or HTTP. Malicious cancel messages (6.3) can cause valid articles to be removed from serving agents. Administrators of such agents SHOULD therefore take steps to verify that they originated from the (apparent) poster, the injector or the moderator of the article, or that in other cases they came from a place that is trusted to work within established policies and customs. Such steps SHOULD include the checking of any digital signatures, or other security devices, that may be provided (see a-7.1). Articles containing Supersedes- headers (a-6.15) are effectively cancel messages, and SHOULD be subject to the same checks. Currently, many sites choose to ignore all cancel messages on account of the difficulty of conducting such checks. Improperly configured serving agents can allow articles posted to moderated groups onto the net without first being approved by the moderator. Injecting agents SHOULD verify that moderated articles were received from one of the entities given in their Approved- headers and/or check any digital signatures that may be provided (see a-7.1). The filename parameter of the Archive-header (a-6.12) can be used to attempt to store archived articles in inappropriate locations. Archiving sites should be suspicious of absolute filename parameters, as opposed to those relative to some location of the archiver's choosing. There may be weaknesses in particular implementations that are subject to malicious exploitation. In particular, it has not been unknown for complete shell scripts to be included within Control- headers. Implementors need to be aware of this. Reading agents should be chary of acting automatically upon MIME objects with an "application" Content-Type that could change the state of that agent, except in contexts where such applications are specifically expected (see a-6.21). Even the Content-Type "text/html" could have unexpected side effects on account of embedded objects, especially embedded executable code or URLs that invoke non-news protocols such as HTTP [RFC 2616]. It is therefore generally recommended that reading agents do not enable the execution of such code (since it is extremely unlikely to have a valid C. H. Lindsey [Page 44] News Article Architecture and Protocols September 2004 application within Netnews) and that they only honour URLs referring to other parts of the same article. Non-printable characters embedded in article bodies may have surprising effects on printers or terminals, notably by reconfiguring them in undesirable ways which may become apparent only after the reading agent has terminated. 8.3. Liability There is a presumption that a poster who sends an article to Usenet intends it to be stored on a multitude of serving agents, and has therefore given permission for it to be copied to that extent. Nevertheless, Usenet is not exempt from the Copyright laws, and it should not be assumed that permission has been given for the article to be copied outside of Usenet, nor for its permanent archiving contrary to any Archive-header that may be present. Posters also need to be aware that they are responsible if they breach Copyright, Libel, Harassment or other restrictions relating to material that they post, and that they may possibly find themselves liable for such breaches in jurisdictions far from their own. Serving agents may also be liable in some jurisdictions, especially if the breach has been explicitly drawn to their attention. Users who are concerned about such matters should seek advice from competent legal authorities. 9. IANA Considerations IANA is requested to register the following media types, described elsewhere in this standard, for use with the Content-Type-header, in the IETF tree in accordance with the procedures set out in [RFC 2048]. application/news-transmission (5.1) application/news-groupinfo (5.3) application/news-checkgroups (5.4) IANA is also requested to change the status of the following media type to "OBSOLETE". message/news (5.2) NOTE: "Application/news-transmission" is an update, with clarification and additional optional parameters, to an existing registration. "Message/rfc822" should now be used in place of the obsoleted "message/news". 10. References [To Do: Split this section into Normative and Informative references. This will probably be delayed until the final draft, for technical reasons.] C. H. Lindsey [Page 45] News Article Architecture and Protocols September 2004 [ANSI X3.4] "American National Standard for Information Systems - Coded Character Sets - 7-Bit American National Standard Code for Information Interchange (7-Bit ASCII)", ANSI X3.4, 1986. [ARTICLE] Charles H. Lindsey, "News Article Format and Transmission", draft-ietf-usefor-article-format-*.txt. [NNTP] Clive D.W. Feather, "Network News Transport Protocol", draft- ietf-nntpext-base-*.txt. [RFC 1036] M. Horton and R. Adams, "Standard for Interchange of USENET Messages", RFC 1036, December 1987. [RFC 2045] N. Freed and N. Borenstein, "Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies", RFC 2045, November 1996. [RFC 2048] N. Freed, J. Klensin, and J. Postel, "Multipurpose Internet Mail Extensions (MIME) Part Four: Registration Procedures", RFC 2048, November 1996. [RFC 2119] S. Bradner, "Key words for use in RFCs to Indicate Requirement Levels", RFC 2119, March 1997. [RFC 2298] R. Fajman, "An Extensible Message Format for Message Disposition Notifications", RFC 2298, March 1998. [RFC 2606] D. Eastlake and A. Panitz, "Reserved Top Level DNS Names", RFC 2606, June 1999. [RFC 2616] R. Fielding, J. Gettys, J. Mogul, H. Frystyk, L. Masinter, P. Leach, and T. Berners-Lee, "Hypertext Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999. [RFC 2822] P. Resnick, "Internet Message Format", RFC 2822, April 2001. [RFC 3864] G. Klyne, M. Nottingham, and J. Mogul, "Registration procedures for message header fields", RFC 3864. [RFC 850] Mark R. Horton, "Standard for interchange of Usenet messages", RFC 850, June 1983. [RFC 976] Mark R. Horton, "UUCP mail interchange format standard", RFC 976, February 1986. [Son-of-1036] Henry Spencer, "News article format and transmission", , June 1994. [USEAGE] draft-ietf-usefor-useage-*.txt. [USEFOR] C. H. Lindsey et al, "News Article Format", draft-ietf- usefor-usefor-format-*.txt. C. H. Lindsey [Page 46] News Article Architecture and Protocols September 2004 [USEPRO] This Standard. 11. Acknowledgements TBD 12. Contact Address Editor Charles. H. Lindsey 5 Clerewood Avenue Heald Green Cheadle Cheshire SK8 3JU United Kingdom Phone: +44 161 436 6131 Email: chl@clw.cs.man.ac.uk [ Working group chair Alexey Melnikov ] Comments on this draft should preferably be sent to the mailing list of the Usenet Format Working Group at ietf-usefor@imc.org. Appendix A.1 - A-News Article Format The obsolete "A News" article format consisted of exactly five lines of header information, followed by the body. For example: Aeagle.642 news.misc cbosgd!mhuxj!mhuxt!eagle!jerry Fri Nov 19 16:14:55 1982 Usenet Etiquette - Please Read body body body The first line consisted of an "A" followed by an article ID (analogous to a message identifier and used for similar purposes). The second line was the list of newsgroups. The third line was the path. The fourth was the date, in the format above (all fields fixed width), resembling an Internet date but not quite the same. The fifth was the subject. C. H. Lindsey [Page 47] News Article Architecture and Protocols September 2004 This format is documented for archeological purposes only. Articles MUST NOT be generated in this format. Appendix A.2 - Early B-News Article Format The obsolete pseudo-Internet article format, used briefly during the transition between the A News format and the modern format, followed the general outline of a MAIL message but with some non-standard headers. For example: From: cbosgd!mhuxj!mhuxt!eagle!jerry (Jerry Schwarz) Newsgroups: news.misc Title: Usenet Etiquette -- Please Read Article-I.D.: eagle.642 Posted: Fri Nov 19 16:14:55 1982 Received: Fri Nov 19 16:59:30 1982 Expires: Mon Jan 1 00:00:00 1990 body body body The From-header contained the information now found in the Path- header, plus possibly the full name now typically found in the From- header. The Title-header contained what is now the Subject-content. The Posted-header contained what is now the Date-content. The Article-I.D.-header contained an article ID, analogous to a message identifier and used for similar purposes. The Newsgroups- and Expires-headers were approximately as now. The Received-header contained the date when the latest relaying agent to process the article first saw it. All dates were in the above format, with all fields fixed width, resembling an Internet date but not quite the same. This format is documented for archeological purposes only. Articles MUST NOT be generated in this format. Appendix A.3 - Obsolete Control Messages This present standard obsoletes certain control messages defined in [RFC 1036] (see 6.5), all of which had the effect of requesting a description of a relaying or serving agent's software, or its peering arrangements with neighbouring sites, to be emailed to the article's reply address. Whilst of some utility when Usenet was much smaller than it is now, they had become no more than a tool for the malicious sending of mailbombs. Moreover, many organizations now consider information about their internal connectivity to be confidential. version sendsys whogets senduuname C. H. Lindsey [Page 48] News Article Architecture and Protocols September 2004 "Version" requested details of the transport software in use at a site. "Sendsys" requested the full list of newsgroups taken, and the peering arrangements. "Who gets" was similar, but restricted to a named newsgroup. "Senduuname" resembled "sendsys" but restricted to the list of peers connected by UUCP. Historically, a checkgroups body consisting of one or two lines, the first of the form "-n newsgroup", caused check-groups to apply to only that single newsgroup. Historically, an article posted to a newsgroup whose name had exactly three components of which the third was "ctl" signified that article was to be taken as a control message. The Subject-header specified the actions, in the same way the Control-header does now. These forms are documented for archeological purposes only; they MUST NO LONGER be used. Appendix B - Notices Intellectual Property The IETF takes no position regarding the validity or scope of any intellectual property or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; neither does it represent that it has made any effort to identify any such rights. Information on the IETF's procedures with respect to rights in standards-track and standards-related documentation can be found in BCP-11. Copies of claims of rights made available for publication and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementors or users of this specification can be obtained from the IETF Secretariat. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights which may cover technology that may be required to practice this standard. Please address the information to the IETF Executive Director. Full Copyright Statement Copyright (C) The Internet Society (2003). All Rights Reserved This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other C. H. Lindsey [Page 49] News Article Architecture and Protocols September 2004 Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Appendix C - Change Log [This Appendix is to be removed prior to final publication.] For version 01 1 Numerous texts describing protocol features related to particular headers in parts of [ARTICLE] which were destined to become part of [USEFOR] have been moved to appropriate locations within section 7 of this document. Such revised texts will be found in sections 7.2.2 Steps 4, 6, 7, 10, 11, 12; 7.2.3 Step 1(b); 7.3 introductory paragraphs, Steps 1, 4, 8, 9, and some final paragraphs; 7.4 introductory and final paragraphs; 7.9.1 Step 5. 2 A section on "Duties of a Reading Agent" (7.8) has been added. 3 Some demotions MUST -> SHOULD -> MAY, as noted in pseudo- comments, have been made or proposed in sections 7.3 7.3 Step 4. 4 Part of the procedure for examining Path-headers by relaying agents has been moved to serving agents, as explained in pseudo-comments in section 7.4. 5 Some renumbering of sections and minor textual clarifications. C. H. Lindsey [Page 50]