[Search] [txt|pdf|bibtex] [Tracker] [Email] [Diff1] [Diff2] [Nits]

Versions: 00 01 02 03 04 05 06                                          
Network Working Group                                         P. Hoffman
Internet-Draft                                            VPN Consortium
Updates: 2223 (if approved)                                      T. Bray
Intended status: Informational                          Sun Microsystems
Expires: April 5, 2009                                   October 2, 2008

                   Using non-ASCII Characters in RFCs

Status of this Memo

   By submitting this Internet-Draft, each author represents that any
   applicable patent or other IPR claims of which he or she is aware
   have been or will be disclosed, and any of which he or she becomes
   aware will be disclosed, in accordance with Section 6 of BCP 79.

   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-

   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

   The list of Internet-Draft Shadow Directories can be accessed at

   This Internet-Draft will expire on April 5, 2009.

Copyright Notice

   Copyright (C) The IETF Trust (2008).


   This document specifies a change to the IETF process in which
   Internet Drafts and RFCs are allowed to contain non-ASCII characters.
   The proposed change is to change the encoding of Internet Drafts and
   RFCs to UTF-8.

Hoffman & Bray            Expires April 5, 2009                 [Page 1]

Internet-Draft              Non-ASCII in RFCs               October 2008

1.  Introduction

   The purpose of this document is to specify a way for the IETF to use
   non-ASCII characters in Internet Drafts and RFCs.

   Various guideline documents in the IETF, notably [RFC2223], specify
   that RFCs must use only the US-ASCII character set.  This restriction
   has historically caused problems, notably:

   o  Names and addresses of authors of IETF documents are misspelled

   o  Names and document titles in references are misspelled

   o  Protocol examples that include non-ASCII characters cannot be
      included straightforwardly

   The first two issues cause real problems for people searching for
   RFCs for particular authors or references that contain non-ASCII
   characters.  For many languages that use Latin characters outside the
   ASCII range, there are not absolute mappings between those non-ASCII
   characters and ASCII equivalents.  A common example is that "a-with-
   umlaut" (U+00E4) may be mapped to "a" or to "ae"; many other mapping
   difficulties exist.

   The third issue reduces the effectiveness of IETF specifications;
   Implementors of protocols which carry textual payloads often
   experience difficulty in achieving interoperability related to the
   use of character sets from around the world.  Specifications which
   can provide concrete examples of such protocol scenarios will be of
   significant benefit to these implementors.

   Now that UTF-8 [RFC3629] is nearly universally available in text-
   editing and display systems, the IETF can eliminate these problems by
   allowing RFCs to use UTF-8.

   This document uses example characters as specified in [RFC5137].  Had
   the recommendations from this document already been implemented, this
   alternate representation would, of course, not be necessary.

   It is important to note that this document does not use RFC 2119
   language (MUST, SHOULD, and so on).  Instead, it lists practices that
   the IETF should consider.  If the ideas in this document are adopted,
   the final list of rules for using UTF-8 in Internet Drafts and RFCs
   would be published by the IAOC.  The authors are open to changing
   this and using 2119-style language if the community prefers it.

Hoffman & Bray            Expires April 5, 2009                 [Page 2]

Internet-Draft              Non-ASCII in RFCs               October 2008

2.  Use of UTF-8 in Internet Drafts and RFCs

   Upon publication of this document as an RFC, all existing RFCs and
   Internet Drafts will be considered to be encoded in UTF-8.  The RFC
   Editor needs to change their processes to publish documents that are
   valid UTF-8.

   Similarly, upon acceptance of this document by the IETF, the IAOC
   should direct the IETF Secretariat to have all Internet Drafts
   encoded in UTF-8.  The Secretariat needs to change their processes to
   publish documents that are valid UTF-8.

2.1.  Limits On the Locations In Which Non-ASCII Text May Be Used

   It is suggested that the IETF Secretariat and RFC Editor limit non-
   ASCII characters to the following:

   o  Names and addresses of authors, used at the top of RFCs and in
      Author Contact sections

   o  Names and document titles used in References sections

   o  Quotations from non-English languages

   o  Protocol examples that show non-ASCII characters, for example in
      Internationalized Domain Names (IDNs), Internationalized Resource
      Identifiers (IRIs), and internationalized email addresses.

2.2.  Allowable Character Repertoire

   UTF-8 is an encoding of the Unicode Character Set and can be used to
   any of its numeric codepoints, from 0 to 0x10FFFF inclusive.
   Specifications encoded in UTF-8 should not contain the encodings of
   certain Unicode codepoints.  The codepoint ranges given in this
   section are inclusive:

   o  The "ASCII control characters" in the ranges U+0000 to U+0008, and
      U+000B to U+001F. These lack either visual representations,
      interoperable semantics, or both.

   o  The Surrogate-block range U+D800 to U+DFFF.  These codepoints do
      not identify characters, but exist to support the UTF-16 encoding.

   o  The ZERO WIDTH NO-BREAK SPACE U+FEFF and its mirror image U+FFFE.

   o  The Private-Use-Area ranges, U+E000 to U+F8FF, U+F0000 to U+FFFFD,
      and U+100000 to U+10FFFD.

Hoffman & Bray            Expires April 5, 2009                 [Page 3]

Internet-Draft              Non-ASCII in RFCs               October 2008

   Specifications encoded in UTF-8 should not contain the encodings of
   Unicode codepoints which are "Compatibility Characters", that is,
   those whose properties include a compatibility decomposition.  Note
   that such characters occur rarely and detecting them requires run-
   time access to the Unicode character database, which may not be
   practical in some situations.

2.3.  Normalization

   Due to the way that Unicode uses combining characters, there are
   sometimes multiple codepoint sequences that denote what, to a human,
   is the same character.  For example, the character "lowercase-a-with-
   accent" can be spelled in two ways: as a single character (U+00E1) or
   as two characters (U+0061 followed by U+0301).  This can present
   problems in searching and rendering.

   The process of standardizing on one of these possibilities is
   referred to as "normalization" and several "normalization forms" are
   defined by the Unicode Consortium.  All UTF-8 text appearing in RFCs
   (but not necessarily Internet Drafts) ought to be normalized using
   Normalization Form C.

2.4.  Author and Employer Names

   Authors can choose how to spell their names and the names of their
   employers in the various parts of Internet Drafts they are writing.
   The spelling at the top of the first page of the document needs to
   match the spelling in the "Authors' Addresses" section near the end
   of the document, but the latter can have alternate spellings to help
   those searching documents by name.  Postal information listed in the
   "Authors' Addresses" section can also use non-ASCII.

   For example, assume that an author whose name is <U+6653><U+4E1C>
   F<U+00E4>ltstr<U+00F6>m has a preferred all-ASCII spelling of
   Xiaodong Faltstrom.  Two expected allowed methods for spelling his
   name would be:

   Network Working Group                                 X. Faltstrom
   Internet-Draft                                           ExampleCo
   . . .
   Author's Address

       Xiaodong Faltstrom (<U+6653><U+4E1C> F<U+00E4>ltstr<U+00F6>m)

       Email: xiaodong.faltstrom@example.com

Hoffman & Bray            Expires April 5, 2009                 [Page 4]

Internet-Draft              Non-ASCII in RFCs               October 2008

   Network Working Group                   X. F<U+00E4>ltstr<U+00F6>m
   Internet-Draft                                           ExampleCo
   . . .
   Author's Address

       <U+6653><U+4E1C> F<U+00E4>ltstr<U+00F6>m (Xiaodong Faltstrom)

       Email: xiaodong.faltstrom@example.com

3.  Security Considerations

   A display program that expects only US-ASCII input may fail when it
   encounters octets outside the US-ASCII range of values.  Such a
   failure may become a security issue.  For example, the program may
   display incorrect results for the input.  More seriously, the program
   may have an internal error that causes it to fail in a security-
   compromising fashion.  Note that such a program is vulnerable to many
   attacks other than just showing IETF documents.

   Someone could insert a UTF-8 host name in an RFC that has visually
   confusing characters.  Another person could copy that host name out
   of the RFC and have it resolve to an unintended DNS name.  This
   scenario seems quite far-fetched, given that tracking the RFC back to
   the author is trivial.

4.  IAOC considerations

   If this document is adopted by the IETF, it will be up to the IAOC to
   have the IETF Secretariat and the RFC Editor implement it.  The IAOC
   needs to consider all of the suggested rules in this document, both
   the positive ones (such as allowing additional characters in some
   parts of Internet Drafts and RFCs) and the negative ones (such as
   disallowing particular characters from being used).  The IAOC might
   want to publish proposed instructions to he IETF Secretariat and the
   RFC Editor and ask for community input on the specific instructions.

5.  Informative References

   [RFC2223]  Postel, J. and J. Reynolds, "Instructions to RFC Authors",
              RFC 2223, October 1997.

   [RFC3629]  Yergeau, F., "UTF-8, a transformation format of ISO
              10646", STD 63, RFC 3629, November 2003.

Hoffman & Bray            Expires April 5, 2009                 [Page 5]

Internet-Draft              Non-ASCII in RFCs               October 2008

   [RFC5137]  Klensin, J., "ASCII Escaping of Unicode Characters",
              BCP 137, RFC 5137, February 2008.

Appendix A.  Arguments Against Changing to UTF-8

   Over more than a decade, the question of changing the encoding of
   RFCs to UTF-8 has come up repeatedly.  Although many people wanted
   the change, various people had different reasons why they felt it was
   a bad idea.  This appendix is a summary of those arguments and an
   explanation of why they are no longer as critical as they were long

A.1.  Difficulty in Displaying

   Some text display systems only know how to display US-ASCII.
   Displaying an RFC that uses non-ASCII characters encoded in UTF-8
   will cause those characters to be unreadable.

   There are, of course, still such display systems, and there always
   will be.  However, the number is dwindling as more software is
   improved to display non-ASCII characters and, in particular, to read
   UTF-8 as an encoding.  Of the systems that can only render US-ASCII,
   only a small subset drop non-ASCII characters: the others show an
   incorrect character in its place.  Thus, the person using such a
   system can often see that there is a problem, and can possibly choose
   to get better display software.

A.2.  Difficulty in Printing

   Some printers can only print a limited set of characters due to the
   fact that they are character-oriented, not graphical.  Such printers
   inherently cannot print characters they do not understand.  Almost
   all such printers print the ASCII characters just fine.

   There are, of course, still such printers, and there always will be.
   However, the number is dwindling as older printers are replaced with
   ones that can print graphics so that now-common text features like
   boldface and italics can be printed.

A.3.  Insufficient Fonts

   Almost no display system that can display text that is encoded with
   UTF-8 can display every character in the Unicode repertoire.  Thus,
   some non-ASCII characters that are included in RFCs will not display

   Virtually every system that can display Unicode knows how to

Hoffman & Bray            Expires April 5, 2009                 [Page 6]

Internet-Draft              Non-ASCII in RFCs               October 2008

   substitute a replacement character for ones that cannot be displayed.
   In fact, most such systems have glyphs for rendering unknown
   characters and different glyphs for rendering known characters for
   which the system has no font.

A.4.  Inability to Search for Non-ASCII Characers

   If authors start using non-ASCII characters in their names and/or
   addresses, people who know the characters but are unfamiliar with the
   user interface on their computers may not be able to enter those
   characters in the search criteria.  For example, some people do not
   know how to enter "u-with-umlaut" in their operating system, even
   though the operating system allows such input.

   This is a valid concern, but one that is orthogonal to whether or not
   RFCs should use these characters.  The alternative (never go to
   UTF-8) simply shifts the problem to forcing the user to guess which
   ASCII-only spelling to use when searching.

Appendix B.  Changes from -02 to -03

   Changed the example name from Frank H<U+00F6>rst to <U+6653><U+4E1C>

   In 2.1, changed "It is suggested that the RFC Editor limit..." to "It
   is suggested that the IETF Secretariat and RFC Editor limit..."

   Made 2.4 match 2.1 by saying that postal addresses can be in UTF-8 as

Authors' Addresses

   Paul Hoffman
   VPN Consortium

   Email: paul.hoffman@vpnc.org

   Tim Bray
   Sun Microsystems

   Email: tbray@textuality.com

Hoffman & Bray            Expires April 5, 2009                 [Page 7]

Internet-Draft              Non-ASCII in RFCs               October 2008

Full Copyright Statement

   Copyright (C) The IETF Trust (2008).

   This document is subject to the rights, licenses and restrictions
   contained in BCP 78, and except as set forth therein, the authors
   retain all their rights.

   This document and the information contained herein are provided on an

Intellectual Property

   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights 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; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat 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 implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at


   Funding for the RFC Editor function is provided by the IETF
   Administrative Support Activity (IASA).

Hoffman & Bray            Expires April 5, 2009                 [Page 8]