The Use of URLs as Meta-Syntax for Core Mail List Commands

and their Transport through Message Header Fields

Status of this Memo

This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.

Copyright Notice

Copyright © The Internet Society (1998). All Rights Reserved.

Abstract

The mailing list command specification header fields are a set of structured fields to be added to email messages sent by email distribution lists. Each field typically contains a URL (usually mailto [RFC2368]) locating the relevant information or performing the command directly. The three core header fields described in this document are List-Help, List-Subscribe, and List-Unsubscribe.

There are three other header fields described here which, although not as widely applicable, will have utility for a sufficient number of mailing lists to justify their formalization here. These are List-Post, List-Owner and List-Archive.

By including these header fields, list servers can make it possible for mail clients to provide automated tools for users to perform list functions. This could take the form of a menu item, push button, or other user interface element. The intent is to simplify the user experience, providing a common interface to the often cryptic and varied mailing list manager commands.

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED",  "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119.

1. Introduction

This is a proposal for additional header fields to be added to email messages sent by email distribution lists. The content of each new field is typically a URL - usually mailto [RFC2368] - which locates the relevant information or performs the command directly. MTAs generating the header fields SHOULD usually include a mailto based command, in addition to any other protocols used, in order to support users who do not have access to non-mail-based protocols.

Implementing these fields will be optional. Significant functionality and convenience can be gained by including them, however. Many list managers, especially as the proposal first gains acceptance, MAY choose to implement only one or two of the fields. The List-Help field is the most useful individual field since it provides an access point to detailed user support information, and accommodates almost all existing list managers command sets. The List-Subscribe and List-Unsubscribe fields are also very useful, but cannot describe some list manager syntaxes at this time (those which require variable substitution). See appendix A.5 for an explanation.

The description of command syntax provided by the fields can be used by mail client applications to provide simplified and consistent user access to email distribution list functions. This could take the form of menu items, push buttons, or other user interface elements. The intent is to simplify the user experience, providing a common interface to the often cryptic and varied mailing list manager commands.

Consideration has been given to avoiding the creation of too many fields, while at the same time avoiding the overloading of individual fields and keeping the syntax clear and simple.

The use of these fields does not remove the requirement to support the -Request command address for mailing lists [RFC2142].

2. The Command Syntax

The list header fields are subject to the encoding and character restrictions for mail headers as described in [RFC822]. Additionally, the URL content is further restricted to the set of URL safe characters [RFC1738].

The contents of the list header fields mostly consist of angle- bracket ('<', '>') enclosed URLs, with internal whitespace being ignored. MTAs MUST NOT insert whitespace within the brackets, but client applications should treat any whitespace, that might be inserted by poorly behaved MTAs, as characters to ignore.

A list of multiple, alternate, URLs MAY be specified by a comma- separated list of angle-bracket enclosed URLs. The URLs have order of preference from left to right. The client application should use the left most protocol that it supports, or knows how to access by a separate application. By this mechanism, protocols like http may be specified while still providing the basic mailto support for those clients who do not have access to non-mail protocols. The client should only use one of the available URLs for a command, using another only if the first one used failed.

The use of URLs allows for the use of the syntax with existing URL supporting applications. As the standard for URLs is extended, the list header fields will gain the benefit of those extensions. Additionally, the use of URLs provides access to multiple transport protocols (such as ftp and http) although it is expected that the "mailto" protocol [RFC2368] will be the focus of most use of the list header fields. Use of non-mailto protocols should be considered in light of those users who do not have access to the specified mechanism (those who only have email - with no web access).

Command syntaxes requiring variable fields to be set by the client (such as including the user's email address within a command) are not supported by this implementation. However, systems using such syntaxes SHOULD still take advantage of the List-Help field to provide the user with detailed instructions as needed or - perhaps more usefully - provide access to some form of structured command interface such as an HTML-based form.

The additional complications of supporting variable fields within the command syntax was determined to be too difficult to support by this protocol and would compromise the likelihood of implementation by software authors.

To allow for future extension, client applications MUST follow the following guidelines for handling the contents of the header fields described in this document:

1. Except where noted for specific fields, if the content of the field (following any leading whitespace, including comments) begins with any character other than the opening angle bracket '<', the field SHOULD be ignored.

2. Any characters following an angle bracket enclosed URL SHOULD be ignored, unless a comma is the first non-whitespace/comment character after the closing angle bracket.

3. If a sub-item (comma-separated item) within the field is not an angle-bracket enclosed URL, the remainder of the field (the current, and all subsequent, sub-items) SHOULD be ignored.

3. The List Header Fields

This document presents header fields which will provide the command syntax description for the 'core' and key secondary functions of most email distribution lists. The fields implemented on a given list SHOULD be included on all messages distributed by the list (including command responses to individual users), and on other messages where the message clearly applies to one distinct list. There MUST be no more than one of each field present in any given message.

These fields MUST only be generated by mailing lists, not end users.

3.1. List-Help

The List-Help field is the most important of the header fields described in this document. It would be acceptable for a list manager to include only this field, since by definition it SHOULD direct the user to complete instructions for all other commands. Typically, the URL specified would request the help file, perhaps incorporating an HTML form for list commands, for the list, and alternatively provide access to an instructive website.

Examples:

     List-Help: <mailto:list@host.com?subject=help> (List Instructions)
     List-Help: <mailto:list-manager@host.com?body=info>
     List-Help: <mailto:list-info@host.com> (Info about the list)
     List-Help: <http://www.host.com/list/>, <mailto:list-info@host.com>
     List-Help: <ftp://ftp.host.com/list.txt> (FTP),
     
         <mailto:list@host.com?subject=help>

3.2. List-Unsubscribe

The List-Unsubscribe field describes the command (preferably using mail) to directly unsubscribe the user (removing them from the list).

Examples:

     List-Unsubscribe: <mailto:list@host.com?subject=unsubscribe>
     List-Unsubscribe: (Use this command to get off the list)
         <mailto:list-manager@host.com?body=unsubscribe%20list>
     List-Unsubscribe: <mailto:list-off@host.com>
     
     List-Unsubscribe: <http://www.host.com/list.cgi?cmd=unsub&lst=list>,
         <mailto:list-request@host.com?subject=unsubscribe>

3.3. List-Subscribe

The List-Subscribe field describes the command (preferably using mail) to directly subscribe the user (request addition to the list).

Examples:

     List-Subscribe: <mailto:list@host.com?subject=subscribe>
     List-Subscribe: <mailto:list-request@host.com?subject=subscribe>
     List-Subscribe: (Use this command to join the list)
         <mailto:list-manager@host.com?body=subscribe%20list>
     List-Subscribe: <mailto:list-on@host.com>
     List-Subscribe: <http://www.host.com/list.cgi?cmd=sub&lst=list>,
         <mailto:list-manager@host.com?body=subscribe%20list>

3.4. List-Post

The List-Post field describes the method for posting to the list. This is typically the address of the list, but MAY be a moderator, or potentially some other form of submission. For the special case of a list that does not allow posting (e.g., an announcements list), the List-Post field may contain the special value "NO".

Examples:

     List-Post: <mailto:list@host.com>
     List-Post: <mailto:moderator@host.com> (Postings are Moderated)
     List-Post: <mailto:moderator@host.com?subject=list%20posting>
     List-Post: NO (posting not allowed on this list)

3.5. List-Owner

The List-Owner field identifies the path to contact a human administrator for the list. The URL MAY contain the address of a administrator for the list, the mail system administrator, or any other person who can handle user contact for the list. There is no need to specify List-Owner if it is the same person as the mail system administrator (postmaster).

Examples:

     List-Owner: <mailto:listmom@host.com> (Contact Person for Help)
     List-Owner: <mailto:grant@foo.bar> (Grant Neufeld)
     List-Owner: <mailto:josh@foo.bar?Subject=list>

3.6. List-Archive

The List-Archive field describes how to access archives for the list.

Examples:

     List-Archive: <mailto:archive@host.com?subject=index%20list>
     List-Archive: <ftp://ftp.host.com/pub/list/archive/>
     List-Archive: <http://www.host.com/list/archive/> (Web Archive)

4. Supporting Nested Lists

A list that is a sublist for another list in a nested mailing list hierarchy will need to modify some of the List- header fields, while leaving others as the parent list set them.

Sublists SHOULD remove the parent list's List-Help, List-Subscribe, List-Unsubscribe and List-Owner fields, and SHOULD insert their own versions of those fields.

If the sublist provides its own archive, it SHOULD replace the List- Archive with its own. Otherwise, it MUST leave the List-Archive field untouched.

Dependant on how postings to the list are handled, the sublist MAY replace the List-Post field. The appropriateness of whether to replace List-Post is left to the determination of the individual list managers. If the intention is that postings should be distributed to all members of the primary list, List-Post should not be changed by a sublist in such a way that postings will be distributed only to members of the sublist.

5. Security Considerations

There are very few new security concerns generated with this proposal. Message headers are an existing standard, designed to easily accommodate new types. There may be concern with multiple fields being inserted or headers being forged, but these are problems inherent in Internet email, not specific to the protocol described in this document. Further, the implications are relatively harmless.

Mail list processors should not allow any user-originated list header fields to pass through to their lists, lest they confuse the user and have the potential to create security problems.

On the client side, there may be some concern with posts or commands being sent in error. It is required that the user have a chance to confirm any action before it is executed. In the case of mailto, it may be appropriate to create the correctly formatted message without sending it, allowing the user to see exactly what is happening and giving the user the opportunity to approve or discard the message before it is sent.

All security considerations for the use of URLs [RFC1738] apply equally to this protocol. Mail client applications should not support list header field URLs which could compromise the security of the user's system. This includes the "file://" URL type which could potentially be used to trigger the execution of a local application on some user systems.

6. Acknowledgements

The numerous participants of the List-Header [5], ListMom-Talk [6], List-Managers and MIDA-Mail mailing lists contributed much to the formation and structure of this document.

Keith Moore <moore@cs.utk.edu> and Christopher Allen <ChristopherA@consensus.com> provided guidance on the standards process.

A. Background Discussion

This proposal arose from discussions started on the ListMom-Talk Discussion List [6]. When the discussion reached a sufficient level, a separate list was formed for discussing this proposal, the List Headers Mail List [5] for deeper discussion. We have included summaries of key issues raised, in order to show some of the alternatives examined and reasons for our decisions.

A.1. Multiple header fields vs. a single header field

Use of a single header field for transporting command meta-syntax was rejected for a number of reasons.

Such a field would require the creation of a new meta-syntax in order to describe the list commands (as opposed to the use of the widely deployed URL syntax which was chosen for this implementation). Every additional layer of complexity and newness reduces the likelihood of actual implementation because it will require additional work to support. Also, by using the existing URL syntax, we can profit from the end users' knowledge of that syntax and ability to use it even if their client applications do not support the list header fields.

Restricting the transport of meta-syntax to the use of a single header field also introduces complications with header field size limitations. Most individual commands can easily be described in a single line, but describing a multitude of commands can take up many lines in the field and runs a greater risk of being modified by an existing server on route.

The client implementation is also easier with multiple fields, since each command can be supported and implemented individually, completely independent of the others. Thus, some list managers or mail clients can choose to implement a subset of the fields based on the specific needs of their individual lists.

Finally, the format described in this document is simple and well recognized, which reduces the chances of errors in implementation and parsing.

A.2. URLs vs. parameter lists

URLs are already an established syntax which is flexible, well- defined, and in wide spread use. As its definition matures and expands, the abilities of the list fields will grow as well, without requiring modification of this proposal. URLs are well prepared to handle future protocols and developments, and can easily describe the different existing access protocols such as mailto, http and ftp.

Many clients already have functionality for recognizing, parsing, and evaluating URLs, either internally or by passing the request to a helper application. This makes implementation easier and more realistic. As an example, this existing support for URL parsing allowed us to add prototype list header functionality to existing mail clients (Eudora and Emailer for the Macintosh) without modifying their source code.

A.3. Why not just create a standard command language?

A standard command language, supported by all email list services, would go a long way to reducing the problems of list access that currently plague existing services. It would reduce the amount of learning required by end users and allow for a number of common support tools to be developed.

However, such standardization does pose problems in the areas of multi-lingual support and the custom needs of individual mailing lists. The development of such a standard is also expected to be met with a slow adoption rate by software developers and list service providers.

These points do not preclude the development of such a standard (in fact, it would suggest that we should start sooner rather than later), but we do need a solution that can be widely supported by the current list services.

We can support most existing list manager command syntaxes without a standard command language. By using URLs, we allow alternate access methods a standard command language probably wouldn't enable, such as web based control.

Finally, client support for a standard command language is not at all clear or necessarily simple to implement. The variety and large number of commands existing today would require complicated user interfaces which could be confusing and difficult to implement. By restricting this proposal to the core functions, the client

implementation is much simpler, which significantly increases the likelihood of implementation (as evidenced by the support already announced by a number of client and server application authors).

A.4. Internationalization

Multilingual support is up to the URL standard. If URLs support it, then the List- header fields support it. This is another advantage of using URLs as the building blocks for the list header fields.

A.5. Variable Substitution

Variables would allow the List- header fields to accommodate nearly every existing list manager. However, it would immeasurably increase the complexity of the entire proposal, and possibly involve redefining the URL standard, or force us to use something more complicated (and hence more difficult to implement) than URLs to describe the command syntax.

Parameters would either have to be mandatory (i.e. the user agent doesn't submit the message if it doesn't know what text to substitute) or you need a way to say "if you know this parameter, add its text here; otherwise, do this" where "this" is either: (a) substitute a constant string, or (b) fail.

The reason you would want a facility like this is because some list server applications insist on having certain parameters like users' names, which the user agent might or might not know. e.g. listserv insists on having a first name and a last name if you supply either one.

Which could lead to something like the UNIX shell syntax, where ${foo-bar} means substitute the value of parameter "foo" if "foo" is defined, else substitute the string "bar". Perhaps $foo would mean "substitute the value of parameter foo if it is defined, else substitute the empty string"

This all seems far too complicated for the gains involved, especially since the use of variables can often be avoided.

The use of variables in the command syntaxes of list services appears to be lessening and does not, in any case, apply to all commands. While the unsubscribe and subscribe command header fields may not be usable by those systems which require the use of variables, the help field will still provide end users with a consistent point of access through which they can get support for their use of the list.

A.6. Why not use a specialized MIME part instead of header fields?

MIME parts were considered, but because most mail clients currently either don't support MIME or are not equipped to handle such specialized parts - such an implementation would result in problems for end users. It is also not as easy for many list servers to implement MIME as it is to implement new header fields.

However, we are looking at the design of a MIME part to more fully describe list command syntax, as well as trying to find ways to get it supported by the applicable software.

A.7. Why include a Subscribe command?

Subscribe and Unsubscribe are the key commands needed by almost every list. Other commands, such as digest mode, are not as widely supported.

Additionally, users who have unsubscribed (before going on vacation, or for whatever other reason) may want to resubscribe to a list. Or, a message may be forwarded/bounced from a subscriber to a non- subscriber. Or, the user may change addresses and want to subscribe from their new address. Having the List-Subscribe field available could certainly help in all these cases.

A.8. The Dangers of Header Bloat

   At what point are there just too many header fields?  It really
   varies on a list by list basis. On some lists, the majority of users
   will never be aware of a field unless the client software provides
   some alternative user interface to it (akin to the Reply-To field).
   On others, the users will often see the header fields of messages and
   would be able to recognize the function of the URLs contained within.

The flexibility afforded by the protocol described in this document (in that the header fields may be individually implemented as deemed appropriate) provides list administrators with sufficient 'room to maneuver' to meet their individual needs.

B. Client Implementation

B.1. Guidelines

For 'mailto' URL based commands, mail client applications may choose to provide specialized feedback (such as presenting a dialog or alert), instead of the actual command email message, asking for command confirmation from the user. The feedback should identify the message destination and command within a more descriptive explanation. For example:

"Do you want to send the unsubscription command 'unsubscribe somelist' to 'somelist-request@some.host.com'? Sending the command will result in your removal from the associated list."

If the user has multiple email addresses supported by the mail client, the client application should prompt the user for which address to use when subscribing or performing some other action where the address to use cannot be specifically determined. When unsubscribing or such, the address that is subscribed should be used, unless that is not known by the application and cannot be determined from the message headers.

B.2. Implementation Options

The following implementation possibilities are suggested here to give some idea as to why these new header fields will be useful, and how they could be supported.

In most cases, it may be helpful to disable the interface for the commands when not applicable to the currently selected message.

B.2.1. Key combinations and command lines

On text based systems which utilize command lines or key combinations, each field could be implemented as a separate command. Thus one combination would subscribe the user, another would unsubscribe, a third request help, etc. The commands would only be available on messages containing the list header fields.

B.2.2. Menu items

On graphical systems which have menus, these commands could take the form of a menu or sub-menu of items. For example, a "Lists" menu might appear when viewing messages containing the header fields, with items named "Subscribe", "Unsubscribe", "Get Help", "Post Message to

List", "Contact List Owner" and "Access List Archive". This menu could be disabled when not applicable to the current message or disappear entirely.

B.2.3. Push Buttons and Pallettes

On graphical window systems, buttons could be placed in the window of the message, a toolbar, or in a floating pallette of their own. Each button could correspond to a command, with names "Subscribe", "Unsubscribe", "Get Help", "Post to List", "List Owner" and "Archive". These buttons or pallettes could be disabled when not applicable to the current message or disappear entirely.

B.2.4 Feedback to the User

If using a dialog interface (or other feedback element) the client application MUST include an option for the user to review (and possibly modify) the message before it is sent. The application may also find it useful to provide a link to more detailed context- sensitive assistance about mail list access in general.

References

   [RFC822] Crocker, D., "Standard for the Format of ARPA
            Internet Text Messages", STD 11, RFC 822, August 1982.
   
   [RFC1738] Berners-Lee, T., Masinter, L., and M. McCahill,
             "Uniform Resource Locators (URL)" RFC 1738, December 1994.
   
   [RFC2142] Crocker, D., "Mailbox Names for Common Services, Roles and
             Functions", RFC 2142, May 1997.
   
   [RFC2368] Hoffman, P., Masinter, L., and J. Zawinski, "The mailto URL
             scheme", RFC 2368, July 1998.
   
   [5] "List-Header" Mail list. list-header@list.nisto.com
       <URL:http://www.nisto.com/listspec/mail/>
       <URL:http://www.nisto.com/listspec/>
   
   [6] "ListMom-Talk" Mail list. listmom-talk@skyweyr.com
       <URL:http://cgi.skyweyr.com/ListMom.Home>

Editors' Addresses

Joshua D. Baer
Box 273
4902 Forbes Avenue
Pittsburgh, PA 15213-3799
USA

EMail:

          josh@skyweyr.com

Grant Neufeld
Calgary, Alberta
Canada

   EMail: grant@acm.org
   Web: http://www.nisto.com/

Full Copyright Statement

Copyright © The Internet Society (1998). 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 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.