idnits 2.17.00 (12 Aug 2021) /tmp/idnits12611/draft-ietf-idnabis-rationale-07.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** The document seems to lack a License Notice according IETF Trust Provisions of 28 Dec 2009, Section 6.b.i or Provisions of 12 Sep 2009 Section 6.b -- however, there's a paragraph with a matching beginning. Boilerplate error? (You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Feb 2009 rather than one of the newer Notices. See https://trustee.ietf.org/license-info/.) Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 927: '... CONTEXTUAL RULE REQUIRED category and...' RFC 2119 keyword, line 1765: '...esented directly in ASCII MUST use the...' RFC 2119 keyword, line 1766: '...sion to A-labels MUST be performed pri...' RFC 2119 keyword, line 2154: '... MUST, SHOULD, or MAY) and small ...' Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document seems to contain a disclaimer for pre-RFC5378 work, and may have content which was first submitted before 10 November 2008. The disclaimer is necessary when there are original authors that you have been unable to contact, or if some do not wish to grant the BCP78 rights to the IETF Trust. If you are able to get all authors (current and original) to grant those rights, you can and should remove the disclaimer; otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (February 24, 2009) is 4833 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ASCII' is defined on line 1906, but no explicit reference was found in the text == Unused Reference: 'RFC1034' is defined on line 1983, but no explicit reference was found in the text == Unused Reference: 'RFC1035' is defined on line 1986, but no explicit reference was found in the text == Unused Reference: 'RFC2673' is defined on line 2008, but no explicit reference was found in the text == Unused Reference: 'Unicode-UTR36' is defined on line 2053, but no explicit reference was found in the text == Outdated reference: draft-ietf-idnabis-defs has been published as RFC 5890 == Outdated reference: draft-ietf-idnabis-tables has been published as RFC 5892 ** Obsolete normative reference: RFC 3490 (Obsoleted by RFC 5890, RFC 5891) -- Obsolete informational reference (is this intentional?): RFC 810 (Obsoleted by RFC 952) -- Obsolete informational reference (is this intentional?): RFC 2535 (Obsoleted by RFC 4033, RFC 4034, RFC 4035) -- Obsolete informational reference (is this intentional?): RFC 2671 (Obsoleted by RFC 6891) -- Obsolete informational reference (is this intentional?): RFC 2673 (Obsoleted by RFC 6891) -- Obsolete informational reference (is this intentional?): RFC 3454 (Obsoleted by RFC 7564) -- Obsolete informational reference (is this intentional?): RFC 3491 (Obsoleted by RFC 5891) Summary: 3 errors (**), 0 flaws (~~), 8 warnings (==), 8 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group J. Klensin 3 Internet-Draft February 24, 2009 4 Intended status: Informational 5 Expires: August 28, 2009 7 Internationalized Domain Names for Applications (IDNA): Background, 8 Explanation, and Rationale 9 draft-ietf-idnabis-rationale-07.txt 11 Status of this Memo 13 This Internet-Draft is submitted to IETF in full conformance with the 14 provisions of BCP 78 and BCP 79. 16 Internet-Drafts are working documents of the Internet Engineering 17 Task Force (IETF), its areas, and its working groups. Note that 18 other groups may also distribute working documents as Internet- 19 Drafts. 21 Internet-Drafts are draft documents valid for a maximum of six months 22 and may be updated, replaced, or obsoleted by other documents at any 23 time. It is inappropriate to use Internet-Drafts as reference 24 material or to cite them other than as "work in progress." 26 The list of current Internet-Drafts can be accessed at 27 http://www.ietf.org/ietf/1id-abstracts.txt. 29 The list of Internet-Draft Shadow Directories can be accessed at 30 http://www.ietf.org/shadow.html. 32 This Internet-Draft will expire on August 28, 2009. 34 Copyright Notice 36 Copyright (c) 2009 IETF Trust and the persons identified as the 37 document authors. All rights reserved. 39 This document is subject to BCP 78 and the IETF Trust's Legal 40 Provisions Relating to IETF Documents in effect on the date of 41 publication of this document (http://trustee.ietf.org/license-info). 42 Please review these documents carefully, as they describe your rights 43 and restrictions with respect to this document. 45 This document may contain material from IETF Documents or IETF 46 Contributions published or made publicly available before November 47 10, 2008. The person(s) controlling the copyright in some of this 48 material may not have granted the IETF Trust the right to allow 49 modifications of such material outside the IETF Standards Process. 50 Without obtaining an adequate license from the person(s) controlling 51 the copyright in such materials, this document may not be modified 52 outside the IETF Standards Process, and derivative works of it may 53 not be created outside the IETF Standards Process, except to format 54 it for publication as an RFC or to translate it into languages other 55 than English. 57 Abstract 59 Several years have passed since the original protocol for 60 Internationalized Domain Names (IDNs) was completed and deployed. 61 During that time, a number of issues have arisen, including the need 62 to update the system to deal with newer versions of Unicode. Some of 63 these issues require tuning of the existing protocols and the tables 64 on which they depend. This document provides an overview of a 65 revised system and provides explanatory material for its components. 67 Table of Contents 69 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 70 1.1. Context and Overview . . . . . . . . . . . . . . . . . . . 5 71 1.2. Discussion Forum . . . . . . . . . . . . . . . . . . . . . 5 72 1.3. Terminology . . . . . . . . . . . . . . . . . . . . . . . 6 73 1.3.1. Documents and Standards . . . . . . . . . . . . . . . 6 74 1.3.2. DNS "Name" Terminology . . . . . . . . . . . . . . . . 6 75 1.3.3. New Terminology and Restrictions . . . . . . . . . . . 7 76 1.4. Objectives . . . . . . . . . . . . . . . . . . . . . . . . 7 77 1.5. Applicability and Function of IDNA . . . . . . . . . . . . 8 78 1.6. Comprehensibility of IDNA Mechanisms and Processing . . . 9 79 2. Processing in IDNA2008 . . . . . . . . . . . . . . . . . . . . 10 80 3. Permitted Characters: An Inclusion List . . . . . . . . . . . 11 81 3.1. A Tiered Model of Permitted Characters and Labels . . . . 11 82 3.1.1. PROTOCOL-VALID . . . . . . . . . . . . . . . . . . . . 12 83 3.1.1.1. Contextual Rules . . . . . . . . . . . . . . . . . 12 84 3.1.1.2. Rules and Their Application . . . . . . . . . . . 13 85 3.1.2. DISALLOWED . . . . . . . . . . . . . . . . . . . . . . 13 86 3.1.3. UNASSIGNED . . . . . . . . . . . . . . . . . . . . . . 14 87 3.2. Registration Policy . . . . . . . . . . . . . . . . . . . 14 88 3.3. Layered Restrictions: Tables, Context, Registration, 89 Applications . . . . . . . . . . . . . . . . . . . . . . . 15 90 4. Issues that Constrain Possible Solutions . . . . . . . . . . . 15 91 4.1. Display and Network Order . . . . . . . . . . . . . . . . 16 92 4.2. Entry and Display in Applications . . . . . . . . . . . . 17 93 4.3. Linguistic Expectations: Ligatures, Digraphs, and 94 Alternate Character Forms . . . . . . . . . . . . . . . . 18 95 4.4. Case Mapping and Related Issues . . . . . . . . . . . . . 20 96 4.5. Right to Left Text . . . . . . . . . . . . . . . . . . . . 21 97 5. IDNs and the Robustness Principle . . . . . . . . . . . . . . 22 98 6. Front-end and User Interface Processing for Lookup . . . . . . 23 99 7. Migration from IDNA2003 and Unicode Version Synchronization . 26 100 7.1. Design Criteria . . . . . . . . . . . . . . . . . . . . . 26 101 7.1.1. General IDNA Validity Criteria . . . . . . . . . . . . 26 102 7.1.2. Labels in Registration . . . . . . . . . . . . . . . . 27 103 7.1.3. Labels in Lookup . . . . . . . . . . . . . . . . . . . 28 104 7.2. Changes in Character Interpretations . . . . . . . . . . . 29 105 7.3. More Flexibility in User Agents . . . . . . . . . . . . . 31 106 7.4. The Question of Prefix Changes . . . . . . . . . . . . . . 32 107 7.4.1. Conditions Requiring a Prefix Change . . . . . . . . . 32 108 7.4.2. Conditions Not Requiring a Prefix Change . . . . . . . 33 109 7.4.3. Implications of Prefix Changes . . . . . . . . . . . . 33 110 7.5. Stringprep Changes and Compatibility . . . . . . . . . . . 34 111 7.6. The Symbol Question . . . . . . . . . . . . . . . . . . . 34 112 7.7. Migration Between Unicode Versions: Unassigned Code 113 Points . . . . . . . . . . . . . . . . . . . . . . . . . . 36 114 7.8. Other Compatibility Issues . . . . . . . . . . . . . . . . 37 116 8. Name Server Considerations . . . . . . . . . . . . . . . . . . 38 117 8.1. Processing Non-ASCII Strings . . . . . . . . . . . . . . . 38 118 8.2. DNSSEC Authentication of IDN Domain Names . . . . . . . . 38 119 8.3. Root and other DNS Server Considerations . . . . . . . . . 39 120 9. Internationalization Considerations . . . . . . . . . . . . . 39 121 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 39 122 10.1. IDNA Character Registry . . . . . . . . . . . . . . . . . 40 123 10.2. IDNA Context Registry . . . . . . . . . . . . . . . . . . 40 124 10.3. IANA Repository of IDN Practices of TLDs . . . . . . . . . 40 125 11. Security Considerations . . . . . . . . . . . . . . . . . . . 40 126 11.1. General Security Issues with IDNA . . . . . . . . . . . . 40 127 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 41 128 13. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 41 129 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 42 130 14.1. Normative References . . . . . . . . . . . . . . . . . . . 42 131 14.2. Informative References . . . . . . . . . . . . . . . . . . 43 132 Appendix A. Change Log . . . . . . . . . . . . . . . . . . . . . 45 133 A.1. Changes between Version -00 and Version -01 of 134 draft-ietf-idnabis-rationale . . . . . . . . . . . . . . . 45 135 A.2. Version -02 . . . . . . . . . . . . . . . . . . . . . . . 45 136 A.3. Version -03 . . . . . . . . . . . . . . . . . . . . . . . 46 137 A.4. Version -04 . . . . . . . . . . . . . . . . . . . . . . . 46 138 A.5. Version -05 . . . . . . . . . . . . . . . . . . . . . . . 47 139 A.6. Version -06 . . . . . . . . . . . . . . . . . . . . . . . 47 140 A.7. Version -07 . . . . . . . . . . . . . . . . . . . . . . . 47 141 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 48 143 1. Introduction 145 1.1. Context and Overview 147 The original standards for Internationalized Domain Names (IDNs) were 148 completed and deployed starting in 2003. Those standards are known 149 as Internationalized Domain Names in Applications (IDNA), taken from 150 the name of the highest level standard within the group, RFC 3490 151 [RFC3490]. After those standards were deployed, a number of issues 152 arose that led to a call for a new version of the IDNA protocol and 153 the associated tables, including a subset of those described in a 154 recent IAB report [RFC4690] and the need to update the system to deal 155 with newer versions of Unicode. This document further explains the 156 issues that have been encountered when they are important to 157 understanding of the revised protocols. It also provides an overview 158 of the new IDNA model and explanatory material for it. Additional 159 explanatory material for the specific components of the proposals 160 appears with the associated documents. 162 This document and the associated ones are written from the 163 perspective of an IDNA-aware user, application, or implementation. 164 While they may reiterate fundamental DNS rules and requirements for 165 the convenience of the reader, they make no attempt to be 166 comprehensive about DNS principles and should not be considered as a 167 substitute for a thorough understanding of the DNS protocols and 168 specifications. 170 A good deal of the background material that appeared in RFC 3490 171 [RFC3490] has been removed from this update. That material is either 172 of historical interest only or has been covered from a more recent 173 perspective in RFC 4690 [RFC4690]. 175 This document is not normative. The information it provides is 176 intended to make the rules, tables, and protocol easier to understand 177 and to provide overview information and suggestions for zone 178 administrators and others who need to make policy, deployment, and 179 similar decisions about IDNs. 181 1.2. Discussion Forum 183 [[ RFC Editor: please remove this section. ]] 185 IDNA2008 is being discussed in the IETF "idnabis" Working Group and 186 on the mailing list idna-update@alvestrand.no 188 1.3. Terminology 190 Terminology that is critical for understanding this document and the 191 rest of the documents that make up IDNA2008, appears in 192 [IDNA2008-Defs]. That document also contains roadmap to the IDNA2008 193 document collection. No attempt should be made to understand this 194 document without the definitions and concepts that appear there. 196 1.3.1. Documents and Standards 198 This document uses the term "IDNA2003" to refer to the set of 199 standards that make up and support the version of IDNA published in 200 2003, i.e., those commonly known as the IDNA base specification 201 [RFC3490], Nameprep [RFC3491], Punycode [RFC3492], and Stringprep 202 [RFC3454]. In this document, those names are used to refer, 203 conceptually, to the individual documents, with the base IDNA 204 specification called just "IDNA". 206 The term "IDNA2008" is used to refer to a new version of IDNA as 207 described in this document and in the documents described in the 208 document listing of [IDNA2008-Defs]. IDNA2008 is not dependent on 209 any of the IDNA2003 specifications other than the one for Punycode 210 encoding. References to "these specifications" or "these documents" 211 are to the entire IDNA2008 set. 213 1.3.2. DNS "Name" Terminology 215 These documents depart from historical DNS terminology and usage in 216 one important respect. Over the years, the community has talked very 217 casually about "names" in the DNS, beginning with calling it "the 218 domain name system". That terminology is fine in the very precise 219 sense that the identifiers of the DNS do provide names for objects 220 and addresses. But, in the context of IDNs, the term has introduced 221 some confusion, confusion that has increased further as people have 222 begun to speak of DNS labels in terms of the words or phrases of 223 various natural languages. 225 Historically, many, perhaps most, of the "names" in the DNS have been 226 mnemonics to identify some particular concept, object, or 227 organization. They are typically derived from, or rooted in, some 228 language because most people think in language-based ways. But, 229 because they are mnemonics, they need not obey the orthographic 230 conventions of any language: it is not a requirement that it be 231 possible for them to be "words". 233 This distinction is important because the reasonable goal of an IDN 234 effort is not to be able to write the great Klingon (or language of 235 one's choice) novel in DNS labels but to be able to form a usefully 236 broad range of mnemonics in ways that are as natural as possible in a 237 very broad range of scripts. 239 1.3.3. New Terminology and Restrictions 241 These documents introduce new terminology, and precise definitions, 242 for the terms "U-label", "A-Label", LDH-label (to which all valid 243 pre-IDNA host names conformed), Reserved-LDH-label (R-LDH-label), XN- 244 label, Fake-A-Label, and Non-Reserved-LDH-label (NR-LDH-label). 246 In addition, the term "putative label" has been adopted to refer to a 247 label that may appear to meet certain definitional constraints but 248 has not yet been sufficiently tested for validity. 250 These definitions are illustrated in Figure 1 of the Definitions 251 Document [IDNA2008-Defs]. R-LDH-labels contain "--" in the third and 252 fourth character from the beginning of the label. In IDNA-aware 253 applications, only a subset of these reserved labels is permitted to 254 be used, namely the A-label subset. A-labels are a subset of the 255 R-LDH-labels that begin with the case-insensitive (?) string "xn--". 256 Labels that bear this prefix but which are not otherwise valid fall 257 into the "Fake-A-label" category. The non-reserved labels (NR-LDH- 258 labels) are implicitly valid since they do not trigger any 259 resemblance to IDNA-landr NR-LDH-labels. 261 The creation of the Reserved-LDH category is required for three 262 reasons: 264 o to prevent confusion with pre-IDNA coding forms; 266 o to permit future extensions that would require changing the 267 prefix, no matter how unlikely those might be (see Section 7.4); 268 and 270 o to reduce the opportunities for attacks via the Punycode encoding 271 algorithm itself. 273 1.4. Objectives 275 The intent of the IDNA revision effort, and hence of this document 276 and the associated ones, is to increase the usability and 277 effectiveness of internationalized domain names (IDNs) while 278 preserving or strengthening the integrity of references that use 279 them. The original "hostname" character definitions (see, e.g., 280 [RFC0810]) struck a balance between the creation of useful mnemonics 281 and the introduction of parsing problems or general confusion in the 282 contexts in which domain names are used. The objective of IDNA2008 283 is to preserve that balance while expanding the character repertoire 284 to include extended versions of Roman-derived scripts and scripts 285 that are not Roman in origin. No work of this sort is able to 286 completely eliminate sources of visual or textual confusion: such 287 confusion is possible even under the original host naming rules where 288 only ASCII characters were permitted. However, through the 289 application of different techniques at different points (see 290 Section 3.3), it should be possible to keep problems to an acceptable 291 minimum. One consequence of this general objective is that the 292 desire of some user or marketing community to use a particular string 293 --whether the reason is to try to write sentences of particular 294 languages in the DNS, to express a facsimile of the symbol for a 295 brand, or for some other purpose-- is not a primary goal within the 296 context of applications in the domain name space. 298 1.5. Applicability and Function of IDNA 300 The IDNA specification solves the problem of extending the repertoire 301 of characters that can be used in domain names to include a large 302 subset of the Unicode repertoire. 304 IDNA does not extend the service offered by DNS to the applications. 305 Instead, the applications (and, by implication, the users) continue 306 to see an exact-match lookup service. Either there is a single 307 exactly-matching name or there is no match. This model has served 308 the existing applications well, but it requires, with or without 309 internationalized domain names, that users know the exact spelling of 310 the domain names that are to be typed into applications such as web 311 browsers and mail user agents. The introduction of the larger 312 repertoire of characters potentially makes the set of misspellings 313 larger, especially given that in some cases the same appearance, for 314 example on a business card, might visually match several Unicode code 315 points or several sequences of code points. 317 The IDNA standard does not require any applications to conform to it, 318 nor does it retroactively change those applications. An application 319 can elect to use IDNA in order to support IDN while maintaining 320 interoperability with existing infrastructure. If an application 321 wants to use non-ASCII characters in domain names, IDNA is the only 322 currently-defined option. Adding IDNA support to an existing 323 application entails changes to the application only, and leaves room 324 for flexibility in front-end processing and more specifically in the 325 user interface (see Section 6). 327 A great deal of the discussion of IDN solutions has focused on 328 transition issues and how IDNs will work in a world where not all of 329 the components have been updated. Proposals that were not chosen by 330 the original IDN Working Group would have depended on updating of 331 user applications, DNS resolvers, and DNS servers in order for a user 332 to apply an internationalized domain name in any form or coding 333 acceptable under that method. While processing must be performed 334 prior to or after access to the DNS, IDNA requires no changes to the 335 DNS protocol or any DNS servers or the resolvers on user's computers. 337 IDNA allows the graceful introduction of IDNs not only by avoiding 338 upgrades to existing infrastructure (such as DNS servers and mail 339 transport agents), but also by allowing some rudimentary use of IDNs 340 in applications by using the ASCII-encoded representation of the 341 labels containing non-ASCII characters. While such names are user- 342 unfriendly to read and type, and hence not optimal for user input, 343 they can be used as a last resort to allow rudimentary IDN usage. 344 For example, they might be the best choice for display if it were 345 known that relevant fonts were not available on the user's computer. 346 In order to allow user-friendly input and output of the IDNs and 347 acceptance of some characters as equivalent to those to be processed 348 according to the protocol, the applications need to be modified to 349 conform to this specification. 351 This version of IDNA uses the Unicode character repertoire, for 352 continuity with the original version of IDNA. 354 1.6. Comprehensibility of IDNA Mechanisms and Processing 356 One of the major goals of this work is to improve the general 357 understanding of how IDNA works and what characters are permitted and 358 what happens to them. Comprehensibility and predictability to users 359 and registrants are themselves important motivations and design goals 360 for this effort. The effort includes some new terminology and a 361 revised and extended model, both covered in this section, and some 362 more specific protocol, processing, and table modifications. Details 363 of the latter appear in other documents (see [IDNA2008-Defs]). 365 Several issues are inherent in the application of IDNs and, indeed, 366 almost any other system that tries to handle international characters 367 and concepts. They range from the apparently trivial --e.g., one 368 cannot display a character for which one does not have a font 369 available locally-- to the more complex and subtle. Many people have 370 observed that internationalization is just a tool to enable effective 371 localization while permitting some global uniformity. Issues of 372 display, of exactly how various strings and characters are entered, 373 and so on are inherently issues about localization and user interface 374 design. 376 A protocol such as IDNA can only assume that such operations as data 377 entry and reconciliation of differences in character forms are 378 possible. It may make some recommendations about how display might 379 work when characters and fonts are not available, but they can only 380 be general recommendations and, because display functions are rarely 381 controlled by the types of applications that would call upon IDNA, 382 will rarely be very effective. 384 However, shifting responsibility for character mapping and other 385 adjustments from the protocol (where it was located in IDNA2003) to 386 the user interface or processing before invoking IDNA raises issues 387 about both what that processing should do and about compatibility for 388 references prepared in an IDNA2003 context. Those issues are 389 discussed in Section 6. 391 Operations for converting between local character sets and normalized 392 Unicode are part of this general set of user interface issues. The 393 conversion is obviously not required at all in a Unicode-native 394 system that maintains all strings in Normalization Form C (NFC). 395 (See [Unicode-UAX15] for precise definitions of NFC and NFKC if 396 needed.) It may, however, involve some complexity in a system that 397 is not Unicode-native, especially if the elements of the local 398 character set do not map exactly and unambiguously into Unicode 399 characters or do so in a way that is not completely stable over time. 400 Perhaps more important, if a label being converted to a local 401 character set contains Unicode characters that have no correspondence 402 in that character set, the application may have to apply special, 403 locally-appropriate, methods to avoid or reduce loss of information. 405 Depending on the system involved, the major difficulty may not lie in 406 the mapping but in accurately identifying the incoming character set 407 and then applying the correct conversion routine. If a local 408 operating system uses one of the ISO 8859 character sets or an 409 extensive national or industrial system such as GB18030 [GB18030] or 410 BIG5 [BIG5], one must correctly identify the character set in use 411 before converting to Unicode even though those character coding 412 systems are substantially or completely Unicode-compatible (i.e., all 413 of the code points in them have an exact and unique mapping to 414 Unicode code points). It may be even more difficult when the 415 character coding system in local use is based on conceptually 416 different assumptions than those used by Unicode about, e.g., font 417 encodings used for publications in some Indic scripts. Those 418 differences may not easily yield unambiguous conversions or 419 interpretations even if each coding system is internally consistent 420 and adequate to represent the local language and script. 422 2. Processing in IDNA2008 424 These specifications separate Domain Name Registration and Lookup in 425 the protocol specification. Doing so reflects current practice in 426 which per-registry restrictions and special processing are applied at 427 registration time but not during lookup. Even more important in the 428 longer term, it facilitates incremental addition of permitted 429 character groups to avoid freezing on one particular version of 430 Unicode. 432 The actual registration and lookup protocols for IDNA2008 are 433 specified in [IDNA2008-Protocol]. 435 3. Permitted Characters: An Inclusion List 437 This section provides an overview of the model used to establish the 438 algorithm and character lists of [IDNA2008-Tables] and describes the 439 names and applicability of the categories used there. Note that the 440 inclusion of a character in the first category group (Section 3.1.1) 441 does not imply that it can be used indiscriminately; some characters 442 are associated with contextual rules that must be applied as well. 444 The information given in this section is provided to make the rules, 445 tables, and protocol easier to understand. The normative generating 446 rules that correspond to this informal discussion appear in 447 [IDNA2008-Tables] and the rules that actually determine what labels 448 can be registered or looked up are in [IDNA2008-Protocol]. 450 3.1. A Tiered Model of Permitted Characters and Labels 452 Moving to an inclusion model requires respecifying the list of 453 characters that are permitted in IDNs. In IDNA2003, the role and 454 utility of characters are independent of context and fixed forever 455 (or until the standard is replaced). Making completely context- 456 independent rules globally has proven impractical because some 457 characters, especially those that are called "Join_Controls" in 458 Unicode, are needed to make reasonable use of some scripts but have 459 no visible effect(s) in others. IDNA2003 prohibited those types of 460 characters entirely. But the restrictions led to a consensus that 461 under some conditions, these "joiner" characters were legitimately 462 needed to allow useful mnemonics for some languages and scripts. The 463 requirement to support those characters but limit their use to very 464 specific contexts was reinforced by the observation that handling of 465 particular characters across the languages that use a script, or the 466 use of similar or identical-looking characters in different scripts, 467 is more complex than many people believed it was several years ago. 469 Independently of the characters chosen (see next subsection), the 470 approach is to divide the characters that appear in Unicode into 471 three categories: 473 3.1.1. PROTOCOL-VALID 475 Characters identified as "PROTOCOL-VALID" (often abbreviated 476 "PVALID") are, in general, permitted by IDNA for all uses in IDNs. 477 Their use may be restricted by rules about the context in which they 478 appear or by other rules that apply to the entire label in which they 479 are to be embedded. For example, any label that contains a character 480 in this category that has a "right-to-left" property must be used in 481 context with the "Bidi" rules (see [IDNA2008-Bidi]). 483 The term "PROTOCOL-VALID" is used to stress the fact that the 484 presence of a character in this category does not imply that a given 485 registry need accept registrations containing any of the characters 486 in the category. Registries are still expected to apply judgment 487 about labels they will accept and to maintain rules consistent with 488 those judgments (see [IDNA2008-Protocol] and Section 3.3). 490 Characters that are placed in the "PROTOCOL-VALID" category are 491 expected to never be removed from it or reclassified. While 492 theoretically characters could be removed from Unicode, such removal 493 would be inconsistent with the Unicode stability principles (see 494 [Unicode51], Appendix F) and hence should never occur. 496 3.1.1.1. Contextual Rules 498 Some characters may be unsuitable for general use in IDNs but 499 necessary for the plausible support of some scripts. The two most 500 commonly-cited examples are the zero-width joiner and non-joiner 501 characters (ZWJ, U+200D and ZWNJ, U+200C), but provisions for 502 unambiguous labels may require that other characters be restricted to 503 particular contexts. For example, the ASCII hyphen is not permitted 504 to start or end a label, whether that label contains non-ASCII 505 characters or not. 507 These characters must not appear in IDNs without additional 508 restrictions, typically because they have no visible consequences in 509 most scripts but affect format or presentation in a few others or 510 because they are combining characters that are safe for use only in 511 conjunction with particular characters or scripts. In order to 512 permit them to be used at all, they are specially identified as 513 "CONTEXTUAL RULE REQUIRED" and, when adequately understood, 514 associated with a rule. In addition, the rule will define whether it 515 is to be applied on lookup as well as registration. A distinction is 516 made between characters that indicate or prohibit joining (known as 517 "CONTEXT-JOINER" or "CONTEXTJ") and other characters requiring 518 contextual treatment ("CONTEXT-OTHER" or "CONTEXTO"). Only the 519 former require full testing at lookup time. 521 3.1.1.2. Rules and Their Application 523 The actual rules may be DEFINED or NULL. If present, they may have 524 values of "True" (character may be used in any position in any 525 label), "False" (character may not be used in any label), or may be a 526 set of procedural rules that specify the context in which the 527 character is permitted. 529 Examples of descriptions of typical rules, stated informally and in 530 English, include "Must follow a character from Script XYZ", "Must 531 occur only if the entire label is in Script ABC", "Must occur only if 532 the previous and subsequent characters have the DFG property". 534 Because it is easier to identify these characters than to know that 535 they are actually needed in IDNs or how to establish exactly the 536 right rules for each one, a rule may have a null value in a given 537 version of the tables. Characters associated with null rules are not 538 permitted to appear in putative labels for either registration or 539 lookup. Of course, a later version of the tables might contain a 540 non-null rule. 542 The description of the syntax of the rules, and the rules themselves, 543 appears in [IDNA2008-Tables]. 545 3.1.2. DISALLOWED 547 Some characters are inappropriate for use in IDNs and are thus 548 excluded for both registration and lookup (i.e., IDNA-conforming 549 applications performing name lookup should verify that these 550 characters are absent; if they are present, the label strings should 551 be rejected rather than converted to A-labels and looked up. Some of 552 these characters are problematic for use in IDNs (such as the 553 FRACTION SLASH character, U+2044), while some of them (such as the 554 various HEART symbols, e.g., U+2665, U+2661, and U+2765, see 555 Section 7.6) simply fall outside the conventions for typical 556 identifiers (basically letters and numbers). 558 Of course, this category would include code points that had been 559 removed entirely from Unicode should such removals ever occur. 561 Characters that are placed in the "DISALLOWED" category are expected 562 to never be removed from it or reclassified. If a character is 563 classified as "DISALLOWED" in error and the error is sufficiently 564 problematic, the only recourse would be either to introduce a new 565 code point into Unicode and classify it as "PROTOCOL-VALID" or for 566 the IETF to accept the considerable costs of an incompatible change 567 and replace the relevant RFC with one containing appropriate 568 exceptions. 570 There is provision for exception cases but, in general, characters 571 are placed into "DISALLOWED" if they fall into one or more of the 572 following groups: 574 o The character is a compatibility equivalent for another character. 575 In slightly more precise Unicode terms, application of 576 normalization method NFKC to the character yields some other 577 character. 579 o The character is an upper-case form or some other form that is 580 mapped to another character by Unicode casefolding. 582 o The character is a symbol or punctuation form or, more generally, 583 something that is not a letter, digit, or a mark that is used to 584 form a letter or digit. 586 3.1.3. UNASSIGNED 588 For convenience in processing and table-building, code points that do 589 not have assigned values in a given version of Unicode are treated as 590 belonging to a special UNASSIGNED category. Such code points are 591 prohibited in labels to be registered or looked up. The category 592 differs from DISALLOWED in that code points are moved out of it by 593 the simple expedient of being assigned in a later version of Unicode 594 (at which point, they are classified into one of the other categories 595 as appropriate). 597 The rationale for restricting the processing of UNASSIGNED characters 598 is simply that if such characters were permitted to be looked up, for 599 example, and were later assigned, but subject to some set of 600 contextual rules, un-updated instances of IDNA-aware software might 601 permit lookup of labels containing the previously-unassigned 602 characters while updated versions of IDNA-aware software might 603 restrict their use in lookup, depending on the contextual rules. It 604 should be clear that under no circumstance should an UNASSIGNED 605 character be permitted in a label to be registered as part of a 606 domain name. 608 3.2. Registration Policy 610 While these recommendations cannot and should not define registry 611 policies, registries should develop and apply additional restrictions 612 as needed to reduce confusion and other problems. For example, it is 613 generally believed that labels containing characters from more than 614 one script are a bad practice although there may be some important 615 exceptions to that principle. Some registries may choose to restrict 616 registrations to characters drawn from a very small number of 617 scripts. For many scripts, the use of variant techniques such as 618 those as described in RFC 3843 [RFC3743] and RFC 4290 [RFC4290], and 619 illustrated for Chinese by the tables described in RFC 4713 [RFC4713] 620 may be helpful in reducing problems that might be perceived by users. 622 In general, users will benefit if registries only permit characters 623 from scripts that are well-understood by the registry or its 624 advisers. If a registry decides to reduce opportunities for 625 confusion by constructing policies that disallow characters used in 626 historic writing systems or characters whose use is restricted to 627 specialized, highly technical contexts, some relevant information may 628 be found in Section 2.4 "Specific Character Adjustments", Table 4 629 "Candidate Characters for Exclusion from Identifiers" of 630 [Unicode-UAX31] and Section 3.1. "General Security Profile for 631 Identifiers" in [Unicode-Security]. 633 It is worth stressing that these principles of policy development and 634 application apply at all levels of the DNS, not only, e.g., TLD or 635 SLD registrations and that even a trivial, "anything permitted that 636 is valid under the protocol" policy is helpful in that it helps users 637 and application developers know what to expect. 639 3.3. Layered Restrictions: Tables, Context, Registration, Applications 641 The essence of the character rules in IDNA2008 is based on the 642 realization that there is no single magic bullet for any of the 643 issues associated with a multiscript DNS. Instead, the 644 specifications define a variety of approaches that, together, 645 constitute multiple lines of defense against ambiguity in identifiers 646 and loss of referential integrity. The actual character tables are 647 the first mechanism, protocol rules about how those characters are 648 applied or restricted in context are the second, and those two in 649 combination constitute the limits of what can be done by a protocol 650 alone. As discussed in the previous section (Section 3.2), 651 registries are expected to restrict what they permit to be 652 registered, devising and using rules that are designed to optimize 653 the balance between confusion and risk on the one hand and maximum 654 expressiveness in mnemonics on the other. 656 In addition, there is an important role for user agents in warning 657 against label forms that appear problematic given their knowledge of 658 local contexts and conventions. Of course, no approach based on 659 naming or identifiers alone can protect against all threats. 661 4. Issues that Constrain Possible Solutions 662 4.1. Display and Network Order 664 The correct treatment of domain names requires a clear distinction 665 between Network Order (the order in which the code points are sent in 666 protocols) and Display Order (the order in which the code points are 667 displayed on a screen or paper). The order of labels in a domain 668 name that contains characters that are normally written right to left 669 is discussed in [IDNA2008-Bidi]. In particular, there are questions 670 about the order in which labels are displayed if left to right and 671 right to left labels are adjacent to each other, especially if there 672 are also multiple consecutive appearances of one of the types. The 673 decision about the display order is ultimately under the control of 674 user agents --including web browsers, mail clients, and the like-- 675 which may be highly localized. Even when formats are specified by 676 protocols, the full composition of an Internationalized Resource 677 Identifier (IRI) [RFC3987] or Internationalized Email address 678 contains elements other than the domain name. For example, IRIs 679 contain protocol identifiers and field delimiter syntax such as 680 "http://" or "mailto:" while email addresses contain the "@" to 681 separate local parts from domain names. User agents are not required 682 to use those protocol-based forms directly but often do so. While 683 display, parsing, and processing within a label is specified by the 684 normative documents in the IDNA2008 collection, the relationship 685 between fully-qualified domain names and internationalized labels is 686 unchanged from the base DNS specifications. Comments in this 687 document about such full domain names are explanatory or examples of 688 what might be done and must not be considered normative. 690 Questions remain about protocol constraints implying that the overall 691 direction of these strings will always be left to right (or right to 692 left) for an IRI or email address, or if they even should conform to 693 such rules. These questions also have several possible answers. 694 Should a domain name abc.def, in which both labels are represented in 695 scripts that are written right to left, be displayed as fed.cba or 696 cba.fed? An IRI for clear text web access would, in network order, 697 begin with "http://" and the characters will appear as 698 "http://abc.def" -- but what does this suggest about the display 699 order? When entering a URI to many browsers, it may be possible to 700 provide only the domain name and leave the "http://" to be filled in 701 by default, assuming no tail (an approach that does not work for 702 protocols other than HTTP or whatever is chosen as the default). The 703 natural display order for the typed domain name on a right to left 704 system is fed.cba. Does this change if a protocol identifier, tail, 705 and the corresponding delimiters are specified? 707 While logic, precedent, and reality suggest that these are questions 708 for user interface design, not IETF protocol specifications, 709 experience in the 1980s and 1990s with mixing systems in which domain 710 name labels were read in network order (left to right) and those in 711 which those labels were read right to left would predict a great deal 712 of confusion, and heuristics that sometimes fail, if each 713 implementation of each application makes its own decisions on these 714 issues. 716 Any version of IDNA, including the current one, must be written in 717 terms of the network (transmission on the wire) order of characters 718 in labels and for the labels in complete (fully-qualified) domain 719 names and must be quite precise about those relationships. While 720 some strong suggestions about display order would be desirable to 721 reduce the chances for inconsistent transcription of domain names 722 from printed form, such suggestions are beyond the scope of these 723 specifications. 725 4.2. Entry and Display in Applications 727 Applications can accept domain names using any character set or sets 728 desired by the application developer, specified by the operating 729 system, or dictated by other constraints, and can display domain 730 names in any character set or character coding system. That is, the 731 IDNA protocol does not affect the interface between users and 732 applications. 734 An IDNA-aware application can accept and display internationalized 735 domain names in two formats: the internationalized character set(s) 736 supported by the application (i.e., an appropriate local 737 representation of a U-label), and as an A-label. Applications may 738 allow the display of A-labels, but are encouraged to not do so except 739 as an interface for special purposes, possibly for debugging, or to 740 cope with display limitations. In general, they should allow, but 741 not encourage, user input of that label form. A-labels are opaque 742 and ugly and malicious variations on them are not easily detected by 743 users. Where possible, they should thus only be exposed to users and 744 in contexts in which they are absolutely needed. Because IDN labels 745 can be rendered either as A-labels or U-labels, the application may 746 reasonably have an option for the user to select the preferred method 747 of display; if it does, rendering the U-label should normally be the 748 default. 750 Domain names are often stored and transported in many places. For 751 example, they are part of documents such as mail messages and web 752 pages. They are transported in many parts of many protocols, such as 753 both the control commands of SMTP and associated the message body 754 parts, and in the headers and the body content in HTTP. It is 755 important to remember that domain names appear both in domain name 756 slots and in the content that is passed over protocols. 758 In protocols and document formats that define how to handle 759 specification or negotiation of charsets, labels can be encoded in 760 any charset allowed by the protocol or document format. If a 761 protocol or document format only allows one charset, the labels must 762 be given in that charset. Of course, not all charsets can properly 763 represent all labels. If a U-label cannot be displayed in its 764 entirety, the only choice (without loss of information) may be to 765 display the A-label. 767 In any place where a protocol or document format allows transmission 768 of the characters in internationalized labels, labels should be 769 transmitted using whatever character encoding and escape mechanism 770 the protocol or document format uses at that place. This provision 771 is intended to prevent situations in which, e.g., UTF-8 domain names 772 appear embedded in text that is otherwise in some other character 773 coding. 775 All protocols that use domain name slots (See Section 2.3.1.6 776 [[anchor12: ?? Verify this]] in [IDNA2008-Defs]) already have the 777 capacity for handling domain names in the ASCII charset. Thus, 778 A-labels can inherently be handled by those protocols. 780 4.3. Linguistic Expectations: Ligatures, Digraphs, and Alternate 781 Character Forms 783 [[anchor13: There is some internal redundancy and repetition in the 784 material in this section. Specific suggestions about to reduce or 785 eliminate redundant text would be appreciated. If no such 786 suggestions are received before -07 is posted, this note will be 787 removed.]] 789 Users often have expectations about character matching or equivalence 790 that are based on their own languages and the orthography of those 791 languages. These expectations may not be consistent with forms or 792 actions that can be naturally accommodated in a character coding 793 system, especially if multiple languages are written using the same 794 script but using different conventions. A Norwegian user might 795 expect a label with the ae-ligature to be treated as the same label 796 as one using the Swedish spelling with a-diaeresis even though 797 applying that mapping to English would be astonishing to users. A 798 user in German might expect a label with an o-umlaut and a label that 799 had "oe" substituted, but was otherwise the same, treated as 800 equivalent even though that substitution would be a clear error in 801 Swedish. A Chinese user might expect automatic matching of 802 Simplified and Traditional Chinese characters, but applying that 803 matching for Korean or Japanese text would create considerable 804 confusion. For that matter, an English user might expect "theater" 805 and "theatre" to match. 807 Related issues arise because there are a number of languages written 808 with alphabetic scripts in which single phonemes are written using 809 two characters, termed a "digraph", for example, the "ph" in 810 "pharmacy" and "telephone". (Note that characters paired in this 811 manner can also appear consecutively without forming a digraph, as in 812 "tophat".) Certain digraphs are normally indicated typographically 813 by setting the two characters closer together than they would be if 814 used consecutively to represent different phonemes. Some digraphs 815 are fully joined as ligatures (strictly designating setting totally 816 without intervening white space, although the term is sometimes 817 applied to close set pairs). An example of this may be seen when the 818 word "encyclopaedia" is set with a U+00E6 LATIN SMALL LIGATURE AE 819 (and some would not consider that word correctly spelled unless the 820 ligature form was used or the "a" was dropped entirely). When these 821 ligature and digraph forms have the same interpretation across all 822 languages that use a given script, application of Unicode 823 normalization generally resolves the differences and causes them to 824 match. When they have different interpretations, any requirements 825 for matching must utilize other methods, presumably at the registry 826 level, or users must be educated to understand that matching will not 827 occur. 829 Difficulties arise from the fact that a given ligature may be a 830 completely optional typographic convenience for representing a 831 digraph in one language (as in the above example with some spelling 832 conventions), while in another language it is a single character that 833 may not always be correctly representable by a two-letter sequence 834 (as in the above example with different spelling conventions). This 835 can be illustrated by many words in the Norwegian language, where the 836 "ae" ligature is the 27th letter of a 29-letter extended Latin 837 alphabet. It is equivalent to the 28th letter of the Swedish 838 alphabet (also containing 29 letters), U+00E4 LATIN SMALL LETTER A 839 WITH DIAERESIS, for which an "ae" cannot be substituted according to 840 current orthographic standards. 842 That character (U+00E4) is also part of the German alphabet where, 843 unlike in the Nordic languages, the two-character sequence "ae" is 844 usually treated as a fully acceptable alternate orthography for the 845 "umlauted a" character. The inverse is however not true, and those 846 two characters cannot necessarily be combined into an "umlauted a". 847 This also applies to another German character, the "umlauted o" 848 (U+00F6 LATIN SMALL LETTER O WITH DIAERESIS) which, for example, 849 cannot be used for writing the name of the author "Goethe". It is 850 also a letter in the Swedish alphabet where, like the "a with 851 diaeresis", it cannot be correctly represented as "oe" and in the 852 Norwegian alphabet, where it is represented, not as "o with 853 diaeresis", but as "slashed o", U+00F8. 855 Some of the ligatures that have explicit code points in Unicode were 856 given special handling in IDNA2003 and now pose additional problems 857 in transition. See Section 7.2. 859 Additional cases with alphabets written right to left are described 860 in Section 4.5. 862 Whether ligatures and digraphs are to be treated as a sequence of 863 characters or as a single standalone one constitute a problem that 864 cannot be resolved solely by operating on scripts. They are, 865 however, a key concern in the IDN context. Their satisfactory 866 resolution will require support in policies set by registries, which 867 therefore need to be particularly mindful not just of this specific 868 issue, but of all other related matters that cannot be dealt with on 869 an exclusively algorithmic and global basis. 871 Just as with the examples of different-looking characters that may be 872 assumed to be the same, it is in general impossible to deal with 873 these situations in a system such as IDNA -- or with Unicode 874 normalization generally -- since determining what to do requires 875 information about the language being used, context, or both. 876 Consequently, these specifications make no attempt to treat these 877 combined characters in any special way. However, their existence 878 provides a prime example of a situation in which a registry that is 879 aware of the language context in which labels are to be registered, 880 and where that language sometimes (or always) treats the two- 881 character sequences as equivalent to the combined form, should give 882 serious consideration to applying a "variant" model [RFC3743] 883 [RFC4290], or to prohibiting registration of one the forms entirely, 884 to reduce the opportunities for user confusion and fraud that would 885 result from the related strings being registered to different 886 parties. 888 [[anchor14: Placeholder: A discussion of the Arabic digit issue 889 should go here once it is resolved in some appropriate way.]] 891 4.4. Case Mapping and Related Issues 893 In the DNS, ASCII letters are stored with their case preserved. 894 Matching during the query process is case-independent, but none of 895 the information that might be represented by choices of case has been 896 lost. That model has been accidentally helpful because, as people 897 have created DNS labels by catenating words (or parts of words) to 898 form labels, case has often been used to distinguish among components 899 and make the labels more memorable. 901 The solution of keeping the characters separate but doing matching 902 independent of case is not feasible with IDNA or any IDNA-like model 903 because the matching would then have to be done on the server rather 904 than have characters mapped on the client. That situation was 905 recognized in IDNA2003 and nothing in these specifications 906 fundamentally changes it or could do so. In IDNA2003, all characters 907 are case-folded and mapped. That results in upper-case characters 908 being mapped to lower-case ones and in some other transformations of 909 alternate forms of characters, especially those that do not have (or 910 did not have) upper-case forms. For example, Greek Final Form Sigma 911 (U+03C2) is mapped to the medial form (U+03C3) and Eszett (German 912 Sharp S, U+00DF) is mapped to "ss". Neither of these mappings is 913 reversible because the upper case of U+03C3 is the Upper Case Sigma 914 (U+03A3) and "ss" is an ASCII string. IDNA2008 permits, at the risk 915 of some incompatibility, slightly more flexibility in this area by 916 avoid case folding and treating these characters as themselves. 917 Approaches to handling that incompatibility are discussed in 918 Section 7.2. Although information is lost in IDNA2003's ToASCII 919 operation so that, in some sense, neither Final Sigma nor Eszett can 920 be represented in an IDN at all, its guarantee of mapping when those 921 characters are used as input can be interpreted as violating one of 922 the conditions discussed in Section 7.4.1 and hence requiring a 923 prefix change. The consensus was to not make a prefix change in 924 spite of this issue. Of course, had a prefix change been made (at 925 the costs discussed in Section 7.4.3) there would have been several 926 options, including, if desired, assignment of the character to the 927 CONTEXTUAL RULE REQUIRED category and requiring that it only be used 928 in carefully-selected contexts. 930 4.5. Right to Left Text 932 In order to be sure that the directionality of right to left text is 933 unambiguous, IDNA2003 required that any label in which right to left 934 characters appear both starts and ends with them, not include any 935 characters with strong left to right properties (which excludes other 936 alphabetic characters but permits European digits), and rejects any 937 other string that contains a right to left character. This is one of 938 the few places where the IDNA algorithms (both in IDNA2003 and in 939 IDAN2008) are required to examine an entire label, not just 940 individual characters. The algorithmic model used in IDNA2003 941 rejects the label when the final character in a right to left string 942 requires a combining mark in order to be correctly represented. 944 That prohibition is not acceptable for writing systems for languages 945 written with consonantal alphabets to which diacritical vocalic 946 systems are applied, and for languages with orthographies derived 947 from them where the combining marks may have different functionality. 948 In both cases the combining marks can be essential components of the 949 orthography. Examples of this are Yiddish, written with an extended 950 Hebrew script, and Dhivehi (the official language of Maldives) which 951 is written in the Thaana script (which is, in turn, derived from the 952 Arabic script). IDNA2008 removes the restriction on final combining 953 characters with a new set of rules for right to left scripts and 954 their characters. Those new rules are specified in [IDNA2008-Bidi]. 956 5. IDNs and the Robustness Principle 958 The model of IDNs described in this document can be seen as a 959 particular instance of the "Robustness Principle" that has been so 960 important to other aspects of Internet protocol design. This 961 principle is often stated as "Be conservative about what you send and 962 liberal in what you accept" (See, e.g., Section 1.2.2 of the 963 applications-layer Host Requirements specification [RFC1123]). For 964 IDNs to work well, not only must the protocol be carefully designed 965 and implemented, but zone administrators (registries) must have and 966 require sensible policies about what is registered -- conservative 967 policies -- and implement and enforce them. 969 Conversely, lookup applications are expected to reject labels that 970 clearly violate global (protocol) rules (no one has ever seriously 971 claimed that being liberal in what is accepted requires being 972 stupid). However, once one gets past such global rules and deals 973 with anything sensitive to script or locale, it is necessary to 974 assume that garbage has not been placed into the DNS, i.e., one must 975 be liberal about what one is willing to look up in the DNS rather 976 than guessing about whether it should have been permitted to be 977 registered. 979 As mentioned elsewhere, if a string cannot be successfully found in 980 the DNS after the lookup processing described here, it makes no 981 difference whether it simply wasn't registered or was prohibited by 982 some rule at the registry. Applications should, however, be 983 sensitive to the fact that, because of the possibility of DNS 984 wildcards, the ability to successfully resolve a name does not 985 guarantee that it was actually registered. 987 If lookup applications, as a user interface (UI) or other local 988 matter, decide to warn about some strings that are valid under the 989 global rules but that they perceive as dangerous, that is their 990 prerogative and we can only hope that the market (and maybe 991 regulators) will reinforce the good choices and discourage the poor 992 ones. In this context, a lookup application that decides a string 993 that is valid under the protocol is dangerous and refuses to look it 994 up is in violation of the protocols; one that is willing to look 995 something up, but warns against it, is exercising a local choice. 997 6. Front-end and User Interface Processing for Lookup 999 Domain names may be identified and processed in many contexts. They 1000 may be typed in by users either by themselves or embedded in an 1001 identifier structured for a particular protocol or class of protocols 1002 such a email addresses, URIs, or IRIs. They may occur in running 1003 text or be processed by one system after being provided in another. 1004 Systems may wish to try to normalize URLs so as to determine (or 1005 guess) whether a reference is valid or two references point to the 1006 same object without actually looking the objects up and comparing 1007 them (that is necessary, not just a choice, for URI types that are 1008 not intended to be resolved). Some of these goals may be more easily 1009 and reliably satisfied than others. While there are strong arguments 1010 for any domain name that is placed "on the wire" -- transmitted 1011 between systems -- to be in the zero-ambiguity forms of A-labels, it 1012 is inevitable that programs that process domain names will encounter 1013 U-labels or variant forms. 1015 One source of such forms will be labels created under IDNA2003 1016 because that protocol allowed labels that were transformed from 1017 native-character format by mapping some characters into others before 1018 conversion into ACE ("xn--...") format. One consequence of the 1019 transformations was that, when the ToUnicode and ToASCII operations 1020 of IDNA2003 were applied, ToUnicode(ToASCII(original-label)) often 1021 did not produce the original label. IDNA2008 explicitly defines 1022 A-labels and U-labels as different forms of the same abstract label, 1023 forms that are stable when conversions are performed between them 1024 (without mappings). A different way of explaining this is that there 1025 are, today, domain names in files on the Internet that use characters 1026 that cannot be represented directly in, or recovered from, (A-label) 1027 domain names but for which interpretations are provided by IDNA2003. 1028 There are two major categories of such characters, those that are 1029 removed by NFKC normalization and those upper-case characters that 1030 are mapped to lower-case (there are also a few characters that are 1031 given special-case mapping treatment in Stringprep, including lower- 1032 case characters that are case-folded into other lower-case characters 1033 or strings). 1035 Other issues in domain name identification and processing arise 1036 because IDNA2003 specified that several other characters be treated 1037 as equivalent to the ASCII period (dot, full stop) character used as 1038 a label separator. If a string that might be a domain name appears 1039 in an arbitrary context (such as running text), it is difficult, even 1040 with only ASCII characters, to know whether an actual domain name (or 1041 a protocol parameter like a URI) is present and where it starts and 1042 ends. When using Unicode, this gets even more difficult if treatment 1043 of certain special characters (like the dot that separates labels in 1044 a domain name) depends on context (e.g., prior knowledge of whether 1045 the string represents a domain name or not). That knowledge is not 1046 available if the primary heuristic for identifying the presence of 1047 domain names in strings depends on the presence of dots separating 1048 groups of characters with no intervening spaces. 1050 As discussed elsewhere in this document, the IDNA2008 model removes 1051 all of these mappings and interpretations, including the equivalence 1052 of different forms of dots, from the protocol, discouraging such 1053 mappings and leaving them, when necessary, to local processing. This 1054 should not be taken to imply that local processing is optional or can 1055 be avoided entirely, even if doing so might have been desirable in a 1056 world without IDNA2003 IDNs in files and archives. Instead, unless 1057 the program context is such that it is known that any IDNs that 1058 appear will contain either U-label or A-label forms, or that other 1059 forms can safely be rejected, some local processing of apparent 1060 domain name strings will be required, both to maintain compatibility 1061 with IDNA2003 and to prevent user astonishment. Such local 1062 processing, while not specified in this document or the associated 1063 ones, will generally take one of two forms: 1065 o Generic Preprocessing. 1066 When the context in which the program or system that processes 1067 domain names operates is global, a reasonable balance must be 1068 found that is sensitive to the broad range of local needs and 1069 assumptions while, at the same time, not sacrificing the needs of 1070 one language, script, or user population to those of another. 1072 For this case, the best practice will usually be to apply NFKC and 1073 case-mapping (or, perhaps better yet, Stringprep itself), plus 1074 dot-mapping where appropriate, to the domain name string prior to 1075 applying IDNA. That practice will not only yield a reasonable 1076 compromise of user experience with protocol requirements but will 1077 be almost completely compatible with the various forms permitted 1078 by IDNA2003. 1080 o Highly Localized Preprocessing. 1081 Unlike the case above, there will be some situations in which 1082 software will be highly localized for a particular environment and 1083 carefully adapted to the expectations of users in that 1084 environment. The many discussions about using the Internet to 1085 preserve and support local cultures suggest that these cases may 1086 be more common in the future than they have been so far. 1088 In these cases, we should avoid trying to tell implementers what 1089 they should accept, if only because they are quite likely (and for 1090 good reason) to ignore us. We would assume that they would map 1091 characters that the intuitions of their users would suggest be 1092 mapped and would hope that they would do that mapping as early as 1093 possible, storing A-label or U-label forms in files and 1094 transporting only those forms between systems. One can imagine 1095 switches about whether some sorts of mappings occur, warnings 1096 before applying them or, in a slightly more extreme version of the 1097 approach taken in Internet Explorer version 7 (IE7), systems that 1098 utterly refuse to handle "strange" characters at all if they 1099 appear in U-label form. None of those local decisions are a 1100 threat to interoperability as long as (i) only U-labels and 1101 A-labels are used in interchange with systems outside the local 1102 environment, (ii) no character that would be valid in a U-label as 1103 itself is mapped to something else, (iii) any local mappings are 1104 applied as a preprocessing step (or, for conversions from U-labels 1105 or A-labels to presentation forms, postprocessing), not as part of 1106 IDNA processing proper, and (iv) appropriate consideration is 1107 given to labels that might have entered the environment in 1108 conformance to IDNA2003. 1110 In either case, it is vital that user interface designs and, where 1111 the interfaces are not sufficient, users, be aware that the only 1112 forms of domain names that this protocol anticipates will resolve 1113 globally or compare equal when crude methods (i.e., those not 1114 conforming to the strict definition of label equivalence given in 1115 [IDNA2008-Defs]) are used are those in which all native-script labels 1116 are in U-label form. Forms that assume mapping will occur, 1117 especially forms that were not valid under IDNA2003, may or may not 1118 function in predictable ways across all implementations. 1120 User interfaces involving Latin-based scripts should take special 1121 care when considering how to handle case mapping because small 1122 differences in label strings may cause behavior that is astonishing 1123 to users. Because case-insensitive comparison is done for ASCII 1124 strings by DNS-servers, an all-ASCII label is treated as case- 1125 insensitive. However, if even one of the characters of that string 1126 is replaced by one that requires the label to be given IDN treatment 1127 (e.g., by adding a diacritical mark), then the label effectively 1128 becomes case-sensitive because only lower-case characters are 1129 permitted in IDNs. This suggests that case mapping for Latin-based 1130 scripts (and possibly other scripts with case distinctions) as a 1131 preprocessing matter in applications may be wise to prevent user 1132 astonishment, but, since all applications may not do this and 1133 ambiguity in transport is not desirable, the that case-dependent 1134 forms should not be stored in files. 1136 The comments above apply only in operations that look up names or 1137 interpret files. There are several reasons why registration 1138 activities should require final names and verification of those names 1139 by the would-be registrant. 1141 7. Migration from IDNA2003 and Unicode Version Synchronization 1143 7.1. Design Criteria 1145 As mentioned above and in RFC 4690, two key goals of the IDNA2008 1146 design are to enable applications to be agnostic about whether they 1147 are being run in environments supporting any Unicode version from 3.2 1148 onward and to permit incrementally adding new characters, character 1149 groups, scripts, and other character collections as they are 1150 incorporated into Unicode, without disruption and, in the long term, 1151 without "heavy" processes such as those involving IETF consensus. 1152 (An IETF consensus process is required by the IDNA2008 specifications 1153 and is expected to be required and used until significant experience 1154 accumulates with IDNA operations and new versions of Unicode.) The 1155 mechanisms that support this are outlined above and elsewhere in the 1156 IDNA2008 document set, but this section reviews them in a context 1157 that may be more helpful to those who need to understand the approach 1158 and make plans for it. 1160 7.1.1. General IDNA Validity Criteria 1162 The general criteria for a putative label, and the collection of 1163 characters that make it up, to be considered IDNA-valid are (the 1164 actual rules are rigorously defined in the "Protocol" and "Tables" 1165 documents): 1167 o The characters are "letters", marks needed to form letters, 1168 numerals, or other code points used to write words in some 1169 language. Symbols, drawing characters, and various notational 1170 characters are intended to be permanently excluded -- some because 1171 they are harmful in URI, IRI, or similar contexts (e.g., 1172 characters that appear to be slashes or other reserved URI 1173 punctuation) and others because there is no evidence that they are 1174 important enough to Internet operations or internationalization to 1175 justify expansion of domain names beyond the general principle of 1176 "letters, digits, and hyphen" and the complexities that would come 1177 with it (additional discussion and rationale for the symbol 1178 decision appears in Section 7.6). 1180 o Other than in very exceptional cases, e.g., where they are needed 1181 to write substantially any word of a given language, punctuation 1182 characters are excluded as well. The fact that a word exists is 1183 not proof that it should be usable in a DNS label and DNS labels 1184 are not expected to be usable for multiple-word phrases (although 1185 they are certainly not prohibited if the conventions and 1186 orthography of a particular language cause that to be possible). 1187 Even for English, very common constructions -- contractions like 1188 "don't" or "it's", names that are written with apostrophes such as 1189 "O'Reilly", or characters for which apostrophes are common 1190 substitutes cannot be represented in DNS labels. Words in English 1191 whose usually-preferred spellings include diacritical marks cannot 1192 be represented under the original hostname rules, but most can be 1193 represented if treated as IDNs. 1195 o Characters that are unassigned (have no character assignment at 1196 all) in the version of Unicode being used by the registry or 1197 application are not permitted, even on lookup. The issues 1198 involved in this decision are discussed in Section 7.7. 1200 o Any character that is mapped to another character by a current 1201 version of NFKC is prohibited as input to IDNA (for either 1202 registration or lookup). With a few exceptions, this principle 1203 excludes any character mapped to another by Nameprep [RFC3491]. 1205 Tables used to identify the characters that are IDNA-valid are 1206 expected to be driven by the principles above, principles that are 1207 specified exactly in [IDNA2008-Tables]). The rules given there are 1208 normative, rather than being just an interpretation of the tables. 1210 7.1.2. Labels in Registration 1212 Anyone entering a label into a DNS zone must properly validate that 1213 label -- i.e., be sure that the criteria for that label are met -- in 1214 order for applications to work as intended. This principle is not 1215 new. For example, since the DNS was first deployed, zone 1216 administrators have been expected to verify that names meet 1217 "hostname" [RFC0952] where necessary for the expected applications. 1218 Later addition of special service location formats [RFC2782] imposed 1219 new requirements on zone administrators for the use of labels that 1220 conform to the requirements of those formats. For zones that will 1221 contain IDNs, support for Unicode version-independence requires 1222 restrictions on all strings placed in the zone. In particular, for 1223 such zones: 1225 o Any label that appears to be an A-label, i.e., any label that 1226 starts in "xn--", must be IDNA-valid, i.e., they must be valid 1227 A-labels, as discussed in Section 2 above. 1229 o The Unicode tables (i.e., tables of code points, character 1230 classes, and properties) and IDNA tables (i.e., tables of 1231 contextual rules such as those that appear in the Tables 1232 document), must be consistent on the systems performing or 1233 validating labels to be registered. Note that this does not 1234 require that tables reflect the latest version of Unicode, only 1235 that all tables used on a given system are consistent with each 1236 other. 1238 Under this model, a registry (or entity communicating with a registry 1239 to accomplish name registrations) will need to update its tables -- 1240 both the Unicode-associated tables and the tables of permitted IDN 1241 characters -- to enable a new script or other set of new characters. 1242 It will not be affected by newer versions of Unicode, or newly- 1243 authorized characters, until and unless it wishes to make those 1244 registrations. The zone administrator is also responsible -- under 1245 the protocol and to registrants and users -- for both checking as 1246 required by the protocol and verification that whatever policies it 1247 develops are complied with, whether those policies are for minimizing 1248 risks due to confusable characters and sequences, for preserving 1249 language or script integrity, or for other purposes. Those checking 1250 and verification procedures are more extensive than those that are is 1251 expected of applications systems that look names up. 1253 Systems looking up or resolving DNS labels, especially IDN DNS 1254 labels, must be able to assume that applicable registration rules 1255 were followed for names entered into the DNS. 1257 7.1.3. Labels in Lookup 1259 Anyone looking up a label in a DNS zone is required to 1261 o Maintain a consistent set of tables, as discussed above. As with 1262 registration, the tables need not reflect the latest version of 1263 Unicode but they must be consistent. 1265 o Validate the characters in labels to be looked up only to the 1266 extent of determining that the U-label does not contain either 1267 code points prohibited by IDNA (categorized as "DISALLOWED") or 1268 code points that are unassigned in its version of Unicode. 1270 o Validate the label itself for conformance with a small number of 1271 whole-label rules, notably verifying that there are no leading 1272 combining marks, that the "bidi" conditions are met if right to 1273 left characters appear, that any required contextual rules are 1274 available and that, if such rules are associated with Joiner 1275 Controls, they are tested. 1277 o Avoid validating other contextual rules about characters, 1278 including mixed-script label prohibitions, although such rules may 1279 be used to influence presentation decisions in the user interface. 1280 [[anchor18: Check this, and all similar statements, against 1281 Protocol when that is finished.]] 1283 By avoiding applying its own interpretation of which labels are valid 1284 as a means of rejecting lookup attempts, the lookup application 1285 becomes less sensitive to version incompatibilities with the 1286 particular zone registry associated with the domain name. 1288 An application or client that processes names according to this 1289 protocol and then resolves them in the DNS will be able to locate any 1290 name that is validly registered, as long as its version of the 1291 Unicode-associated tables is sufficiently up-to-date to interpret all 1292 of the characters in the label. Messages to users should distinguish 1293 between "label contains an unallocated code point" and other types of 1294 lookup failures. A failure on the basis of an old version of Unicode 1295 may lead the user to a desire to upgrade to a newer version, but will 1296 have no other ill effects (this is consistent with behavior in the 1297 transition to the DNS when some hosts could not yet handle some forms 1298 of names or record types). 1300 7.2. Changes in Character Interpretations 1302 [[anchor19: Note in Draft: This subsection is completely new in 1303 version -04 and has been further tuned in -05 and -06 of this 1304 document. It could almost certainly use improvement, although this 1305 note will be removed if there are not significant suggestions about 1306 the -06 version. It also contains some material that is redundant 1307 with material in other sections. I have not tried to remove that 1308 material and will not do so until the WG concludes that this section 1309 is relatively stable, but would appreciate help in identifying what 1310 should be removed or how this might be enhanced to contain more of 1311 that other material. --JcK]] 1313 In those scripts that make case distinctions, there are a few 1314 characters for which an obvious and unique upper case character has 1315 not historically been available to match a lower case one or vice 1316 versa. For those characters, the mappings used in constructing the 1317 Stringprep tables for IDNA2003, performed using the Unicode CaseFold 1318 operation (See Section 5.8 of the Unicode Standard [Unicode51]), 1319 generate different characters or sets of characters. Those 1320 operations are not reversible and lose even more information than 1321 traditional upper case or lower case transformations, but are more 1322 useful than those transformations for comparison purposes. Two 1323 notable characters of this type are the German character Eszett 1324 (Sharp S, U+00DF) and the Greek Final Form Sigma (U+03C2). The 1325 former is case-folded to the ASCII string "ss", the latter to a 1326 medial (Lower Case) Sigma (U+03C3). 1328 The decision to eliminate mappings, including case folding, from the 1329 IDNA2008 protocol in order to make A-labels and U-labels idempotent 1330 made these characters problematic. If they were to be disallowed, 1331 important words and mnemonics could not be written in 1332 orthographically reasonable ways. If they were to be permitted as 1333 characters distinct from the forms produced by case folding, there 1334 would be no information loss and registries would have maximum 1335 flexibility, but labels using those characters that were looked up 1336 according to IDNA2003 rules would be transformed into A-labels using 1337 their case-mapped variations while lookup according to IDNA2008 rules 1338 would be based on different A-labels that represented the actual 1339 characters. 1341 With the understanding that there would be incompatibility either way 1342 but a judgment that the incompatibility was not significant enough to 1343 just a prefix change, the WG concluded that Eszett and Final Form 1344 Sigma should be treated as distinct and Protocol-Valid characters. 1346 The decision faces registries, especially registries maintaining 1347 zones for third parties, with a variation on what has become a 1348 familiar problem: how to introduce a new service in a way that does 1349 not create confusion or significantly weaken or invalidate existing 1350 identifiers. 1352 There have traditionally been several approaches to problems of this 1353 type. Without any preference or claim to completeness, these are: 1355 o Do not permit use of the newly-available character at the registry 1356 level. This might cause lookup failures if a domain name were to 1357 be written with the expectation of the IDNA2003 mapping behavior, 1358 but would eliminate any possibility of false matches. 1360 o Hold a "sunrise"-like arrangement in which holders of the 1361 previously-mapped labels (labels containing "ss" in the Eszett 1362 case or ones containing Lower Case Sigma in the Final Sigma case) 1363 are given priority (and perhaps other benefits) for registering 1364 the corresponding string containing the newly-available 1365 characters. 1367 o Adopt some sort of "variant" approach in which registrants either 1368 obtained labels with both character forms or one of them was 1369 blocked from registration by anyone but the registrant of the 1370 other form. 1372 In principle, lookup applications could also compensate for the 1373 difference in interpretation by looking up the string according to 1374 the interpretation specified in these documents and then, if that 1375 failed, doing the lookup with the mapping, simulating the IDNA2003 1376 interpretation. The risk of false positives is such that this is 1377 generally to be discouraged unless the application is able to engage 1378 in a "is this what you meant" dialogue with the end user. 1380 7.3. More Flexibility in User Agents 1382 These documents do not specify mappings between one character or code 1383 point and others for any reason. Instead, they prohibit the 1384 characters that would be mapped to others by normalization, upper 1385 case to lower case changes, or other rules. As examples, while 1386 mathematical characters based on Latin ones are accepted as input to 1387 IDNA2003, they are prohibited in IDNA2008. Similarly, double-width 1388 characters and other variations are prohibited as IDNA input. 1390 Since the rules in [IDNA2008-Tables] have the effect that only 1391 strings that are not transformed by NFKC are valid, if an application 1392 chooses to perform NFKC normalization before lookup, that operation 1393 is safe since this will never make the application unable to look up 1394 any valid string. However, as discussed above, the application 1395 cannot guarantee that any other application will perform that 1396 mapping, so it should be used only with caution and for informed 1397 users. 1399 In many cases these prohibitions should have no effect on what the 1400 user can type as input to the lookup process. It is perfectly 1401 reasonable for systems that support user interfaces to perform some 1402 character mapping that is appropriate to the local environment. This 1403 would normally be done prior to actual invocation of IDNA. At least 1404 conceptually, the mapping would be part of the Unicode conversions 1405 discussed above and in [IDNA2008-Protocol]. However, those changes 1406 will be local ones only -- local to environments in which users will 1407 clearly understand that the character forms are equivalent. For use 1408 in interchange among systems, it appears to be much more important 1409 that U-labels and A-labels can be mapped back and forth without loss 1410 of information. 1412 One specific, and very important, instance of this strategy arises 1413 with case-folding. In the ASCII-only DNS, names are looked up and 1414 matched in a case-independent way, but no actual case-folding occurs. 1415 Names can be placed in the DNS in either upper or lower case form (or 1416 any mixture of them) and that form is preserved, returned in queries, 1417 and so on. IDNA2003 simulated that behavior for non-ASCII strings by 1418 performing case-folding at registration time (resulting in only 1419 lower-case IDNs in the DNS) and when names were looked up. 1421 As suggested earlier in this section, it appears to be desirable to 1422 do as little character mapping as possible consistent with having 1423 Unicode work correctly (e.g., NFC mapping to resolve different 1424 codings for the same character is still necessary although the 1425 specifications require that it be performed prior to invoking the 1426 protocol) and to make the mapping between A-labels and U-labels 1427 idempotent. Case-mapping is not an exception to this principle. If 1428 only lower case characters can be registered in the DNS (i.e., be 1429 present in a U-label), then IDNA2008 should prohibit upper-case 1430 characters as input (and therefore does so). Some other 1431 considerations reinforce this conclusion. For example, an essential 1432 element of the ASCII case-mapping functions is that, for individual 1433 characters, uppercase(character) must be equal to 1434 uppercase(lowercase(character)). That requirement may not be 1435 satisfied with IDNs. For example, there are some characters in 1436 scripts that use case distinction that do not have counterparts in 1437 one case or the other. The relationship between upper case and lower 1438 case may even be language-dependent, with different languages (or 1439 even the same language in different areas) expecting different 1440 mappings. Of course, the expectations of users who are accustomed to 1441 a case-insensitive DNS environment will probably be well-served if 1442 user agents perform case folding prior to IDNA processing, but the 1443 IDNA procedures themselves should neither require such mapping nor 1444 expect them when they are not natural to the localized environment. 1446 7.4. The Question of Prefix Changes 1448 The conditions that would require a change in the IDNA ACE prefix 1449 ("xn--" for the version of IDNA specified in [RFC3490]) have been a 1450 great concern to the community. A prefix change would clearly be 1451 necessary if the algorithms were modified in a manner that would 1452 create serious ambiguities during subsequent transition in 1453 registrations. This section summarizes our conclusions about the 1454 conditions under which changes in prefix would be necessary and the 1455 implications of such a change. 1457 7.4.1. Conditions Requiring a Prefix Change 1459 An IDN prefix change is needed if a given string would be looked up 1460 or otherwise interpreted differently depending on the version of the 1461 protocol or tables being used. Consequently, work to update IDNs 1462 would require a prefix change if, and only if, one of the following 1463 four conditions were met: 1465 1. The conversion of an A-label to Unicode (i.e., a U-label) yields 1466 one string under IDNA2003 (RFC3490) and a different string under 1467 IDNA2008. 1469 2. An input string that is valid under IDNA2003 and also valid under 1470 IDNA2008 yields two different A-labels with the different 1471 versions of IDNA. This condition is believed to be essentially 1472 equivalent to the one above except for a very small number of 1473 edge cases which may not, pragmatically, justify a prefix change 1474 (See Section 7.2). 1476 Note, however, that if the input string is valid under one 1477 version and not valid under the other, this condition does not 1478 apply. See the first item in Section 7.4.2, below. 1480 3. A fundamental change is made to the semantics of the string that 1481 is inserted in the DNS, e.g., if a decision were made to try to 1482 include language or specific script information in that string, 1483 rather than having it be just a string of characters. 1485 4. A sufficiently large number of characters is added to Unicode so 1486 that the Punycode mechanism for block offsets no longer has 1487 enough capacity to reference the higher-numbered planes and 1488 blocks. This condition is unlikely even in the long term and 1489 certain not to arise in the next few years. 1491 7.4.2. Conditions Not Requiring a Prefix Change 1493 In particular, as a result of the principles described above, none of 1494 the following changes require a new prefix: 1496 1. Prohibition of some characters as input to IDNA. This may make 1497 names that are now registered inaccessible, but does not require 1498 a prefix change. 1500 2. Adjustments in IDNA tables or actions, including normalization 1501 definitions, that affect characters that were already invalid 1502 under IDNA2003. 1504 3. Changes in the style of the IDNA definition that does not alter 1505 the actions performed by IDNA. 1507 7.4.3. Implications of Prefix Changes 1509 While it might be possible to make a prefix change, the costs of such 1510 a change are considerable. Even if they wanted to do so, registries 1511 could not convert all IDNA2003 ("xn--") registrations to a new form 1512 at the same time and synchronize that change with applications 1513 supporting lookup. Unless all existing registrations were simply to 1514 be declared invalid (and perhaps even then) systems that needed to 1515 support both labels with old prefixes and labels with new ones would 1516 first process a putative label under the IDNA2008 rules and try to 1517 look it up and then, if it were not found, would process the label 1518 under IDNA2003 rules and look it up again. That process could 1519 significantly slow down all processing that involved IDNs in the DNS 1520 especially since, in principle, a fully-qualified name could contain 1521 a mixture of labels that were registered with the old and new 1522 prefixes, a situation that would make the use of DNS caching very 1523 difficult. In addition, looking up the same input string as two 1524 separate A-labels would create some potential for confusion and 1525 attacks, since they could, in principle, map to different targets and 1526 then resolve to different entries in the DNS. 1528 Consequently, a prefix change is to be avoided if at all possible, 1529 even if it means accepting some IDNA2003 decisions about character 1530 distinctions as irreversible and/or giving special treatment to edge 1531 cases. 1533 7.5. Stringprep Changes and Compatibility 1535 The Nameprep [RFC3491] specification, a key part of IDNA2003, is a 1536 profile of Stringprep [RFC3454]. While Nameprep is a Stringprep 1537 profile specific to IDNA, Stringprep is used by a number of other 1538 protocols. Concerns have been expressed about problems for non-DNS 1539 uses of Stringprep being caused by changes to the specification 1540 intended to improve the handling of IDNs, most notably as this might 1541 affect identification and authentication protocols. The proposed new 1542 inclusion tables [IDNA2008-Tables], the reduction in the number of 1543 characters permitted as input for registration or lookup (Section 3), 1544 and even the proposed changes in handling of right to left strings 1545 [IDNA2008-Bidi] either give interpretations to strings prohibited 1546 under IDNA2003 or prohibit strings that IDNA2003 permitted. The 1547 IDNA2008 protocol does not use either Nameprep or Stringprep at all, 1548 so there are no side-effect changes to other protocols. 1550 It is particularly important to keep IDNA processing separate from 1551 processing for various security protocols because some of the 1552 constraints that are necessary for smooth and comprehensible use of 1553 IDNs may be unwanted or undesirable in other contexts. For example, 1554 the criteria for good passwords or passphrases are very different 1555 from those for desirable IDNs: passwords should be hard to guess, 1556 while domain names should normally be easily memorable. Similarly, 1557 internationalized SCSI identifiers and other protocol components are 1558 likely to have different requirements than IDNs. 1560 7.6. The Symbol Question 1562 One of the major differences between this specification and the 1563 original version of IDNA is that the original version permitted non- 1564 letter symbols of various sorts, including punctuation and line- 1565 drawing symbols, in the protocol. They were always discouraged in 1566 practice. In particular, both the "IESG Statement" about IDNA and 1567 all versions of the ICANN Guidelines specify that only language 1568 characters be used in labels. This specification disallows symbols 1569 entirely. There are several reasons for this, which include: 1571 o As discussed elsewhere, the original IDNA specification assumed 1572 that as many Unicode characters as possible should be permitted, 1573 directly or via mapping to other characters, in IDNs. This 1574 specification operates on an inclusion model, extrapolating from 1575 the LDH rules -- which have served the Internet very well -- to a 1576 Unicode base rather than an ASCII base. 1578 o Most Unicode names for letters are, in most cases, fairly 1579 intuitive, unambiguous and recognizable to users of the relevant 1580 script. Symbol names are more problematic because there may be no 1581 general agreement on whether a particular glyph matches a symbol; 1582 there are no uniform conventions for naming; variations such as 1583 outline, solid, and shaded forms may or may not exist; and so on. 1584 As just one example, consider a "heart" symbol as it might appear 1585 in a logo that might be read as "I love...". While the user might 1586 read such a logo as "I love..." or "I heart...", considerable 1587 knowledge of the coding distinctions made in Unicode is needed to 1588 know that there more than one "heart" character (e.g., U+2665, 1589 U+2661, and U+2765) and how to describe it. These issues are of 1590 particular importance if strings are expected to be understood or 1591 transcribed by the listener after being read out loud. 1592 [[anchor20: The above paragraph remains controversial as to 1593 whether it is valid. The WG will need to make a decision if this 1594 section is not dropped entirely.]] 1596 o Consider the case of a screen reader used by blind Internet users 1597 who must listen to renderings of IDN domain names and possibly 1598 reproduce them on the keyboard. 1600 o As a simplified example of this, assume one wanted to use a 1601 "heart" or "star" symbol in a label. This is problematic because 1602 those names are ambiguous in the Unicode system of naming (the 1603 actual Unicode names require far more qualification). A user or 1604 would-be registrant has no way to know -- absent careful study of 1605 the code tables -- whether it is ambiguous (e.g., where there are 1606 multiple "heart" characters) or not. Conversely, the user seeing 1607 the hypothetical label doesn't know whether to read it -- try to 1608 transmit it to a colleague by voice -- as "heart", as "love", as 1609 "black heart", or as any of the other examples below. 1611 o The actual situation is even worse than this. There is no 1612 possible way for a normal, casual, user to tell the difference 1613 between the hearts of U+2665 and U+2765 and the stars of U+2606 1614 and U+2729 or the without somehow knowing to look for a 1615 distinction. We have a white heart (U+2661) and few black hearts. 1616 Consequently, describing a label as containing a heart hopelessly 1617 ambiguous: we can only know that it contains one of several 1618 characters that look like hearts or have "heart" in their names. 1620 In cities where "Square" is a popular part of a location name, one 1621 might well want to use a square symbol in a label as well and 1622 there are far more squares of various flavors in Unicode than 1623 there are hearts or stars. 1625 o The consequence of these ambiguities of description and 1626 dependencies on distinctions that were, or were not, made in 1627 Unicode codings is that symbols are a very poor basis for reliable 1628 communication. Consistent with this conclusion, the Unicode 1629 standard recommends that strings used in identifiers not contain 1630 symbols or punctuation [Unicode-UAX31]. Of course, these 1631 difficulties with symbols do not arise with actual pictographic 1632 languages and scripts which would be treated like any other 1633 language characters; the two should not be confused. 1635 7.7. Migration Between Unicode Versions: Unassigned Code Points 1637 In IDNA2003, labels containing unassigned code points are looked up 1638 on the assumption that, if they appear in labels and can be mapped 1639 and then resolved, the relevant standards must have changed and the 1640 registry has properly allocated only assigned values. 1642 In the protocol as described in these documents, strings containing 1643 unassigned code points must not be either looked up or registered. 1644 There are several reasons for this, with the most important ones 1645 being: 1647 o It cannot be known in advance, and with sufficient reliability, 1648 that a code point that was not previously assigned will not be 1649 assigned to a compatibility character or one that would be 1650 otherwise disallowed by the rules in [IDNA2008-Tables]. In 1651 IDNA2003, since there is no direct dependency on NFKC 1652 (Stringprep's tables are based on NFKC, but IDNA2003 depends only 1653 on Stringprep), allocation of a compatibility character might 1654 produce some odd situations, but it would not be a problem. In 1655 IDNA2008, where compatibility characters are assigned to 1656 DISALLOWED unless character-specific exceptions are made, 1657 permitting strings containing unassigned characters to be looked 1658 up would permit violating the principle that characters in 1659 DISALLOWED are not looked up. 1661 o The Unicode Standard specifies that an unassigned code point 1662 normalizes (and, where relevant, case folds) to itself. If the 1663 code point is later assigned to a character, and particularly if 1664 the newly-assigned code point has a combining class that 1665 determines its placement relative to other combining characters, 1666 it could normalize to some other code point or sequence, creating 1667 confusion and/or violating other rules listed here. 1669 o Tests involving the context of characters (e.g., some characters 1670 being permitted only adjacent to ones of specific types but 1671 otherwise invisible or very problematic for other reasons) and 1672 integrity tests on complete labels are needed. Unassigned code 1673 points cannot be permitted because one cannot determine whether 1674 particular code points will require contextual rules (and what 1675 those rules should be) before characters are assigned to them and 1676 the properties of those characters fully understood. 1678 o More generally, the status of an unassigned character with regard 1679 to the DISALLOWED and PROTOCOL-VALID categories, and whether 1680 contextual rules are required with the latter, cannot be evaluated 1681 until a character is actually assigned and known. By contrast, 1682 characters that are actually DISALLOWED are placed in that 1683 category only as a consequence of rules applied to known 1684 properties or per-character evaluation. 1686 Another way to look at this is that permitting an unassigned 1687 character to be looked up is nearly equivalent to reclassifying a 1688 character from DISALLOWED to PROTOCOL-VALID since different systems 1689 will interpret the character in different ways. 1691 It is possible to argue that the issues above are not important and 1692 that, as a consequence, it is better to retain the principle of 1693 looking up labels even if they contain unassigned characters because 1694 all of the important scripts and characters have been coded as of 1695 Unicode 5.1 and hence unassigned code points will be assigned only to 1696 obscure characters or archaic scripts. Unfortunately, that does not 1697 appear to be a safe assumption for at least two reasons. First, much 1698 the same claim of completeness has been made for earlier versions of 1699 Unicode. The reality is that a script that is obscure to much of the 1700 world may still be very important to those who use it. Cultural and 1701 linguistic preservation principles make it inappropriate to declare 1702 the script of no importance in IDNs. Second, we already have 1703 counterexamples in, e.g., the relationships associated with new Han 1704 characters being added (whether in the BMP or in Unicode Plane 2). 1706 7.8. Other Compatibility Issues 1708 The 2003 IDNA model includes several odd artifacts of the context in 1709 which it was developed. Many, if not all, of these are potential 1710 avenues for exploits, especially if the registration process permits 1711 "source" names (names that have not been processed through IDNA and 1712 Nameprep) to be registered. As one example, since the character 1713 Eszett, used in German, is mapped by IDNA2003 into the sequence "ss" 1714 rather than being retained as itself or prohibited, a string 1715 containing that character but that is otherwise in ASCII is not 1716 really an IDN (in the U-label sense defined above) at all. After 1717 Nameprep maps the Eszett out, the result is an ASCII string and so 1718 does not get an xn-- prefix, but the string that can be displayed to 1719 a user appears to be an IDN. The newer version of the protocol 1720 eliminates this artifact. A character is either permitted as itself 1721 or it is prohibited; special cases that make sense only in a 1722 particular linguistic or cultural context can be dealt with as 1723 localization matters where appropriate. 1725 8. Name Server Considerations 1727 8.1. Processing Non-ASCII Strings 1729 Existing DNS servers do not know the IDNA rules for handling non- 1730 ASCII forms of IDNs, and therefore need to be shielded from them. 1731 All existing channels through which names can enter a DNS server 1732 database (for example, master files (as described in RFC 1034) and 1733 DNS update messages [RFC2136]) are IDN-unaware because they predate 1734 IDNA. Other sections of this document provide the needed shielding 1735 by ensuring that internationalized domain names entering DNS server 1736 databases through such channels have already been converted to their 1737 equivalent ASCII A-label forms. 1739 Because of the distinction made between the algorithms for 1740 Registration and Lookup in [IDNA2008-Protocol] (a domain name 1741 containing only ASCII codepoints can not be converted to an A-label), 1742 there can not be more than one A-label form for any given U-label. 1744 As specified in RFC 2181 [RFC2181], the DNS protocol explicitly 1745 allows domain labels to contain octets beyond the ASCII range 1746 (0000..007F), and this document does not change that. Note, however, 1747 that there is no defined interpretation of octets 0080..00FF as 1748 characters. If labels containing these octets are returned to 1749 applications, unpredictable behavior could result. The A-label form, 1750 which cannot contain those characters, is the only standard 1751 representation for internationalized labels in the DNS protocol. 1753 8.2. DNSSEC Authentication of IDN Domain Names 1755 DNS Security (DNSSEC) [RFC2535] is a method for supplying 1756 cryptographic verification information along with DNS messages. 1757 Public Key Cryptography is used in conjunction with digital 1758 signatures to provide a means for a requester of domain information 1759 to authenticate the source of the data. This ensures that it can be 1760 traced back to a trusted source, either directly or via a chain of 1761 trust linking the source of the information to the top of the DNS 1762 hierarchy. 1764 IDNA specifies that all internationalized domain names served by DNS 1765 servers that cannot be represented directly in ASCII MUST use the 1766 A-label form. Conversion to A-labels MUST be performed prior to a 1767 zone being signed by the private key for that zone. Because of this 1768 ordering, it is important to recognize that DNSSEC authenticates a 1769 domain name containing A-labels or conventional LDH-labels, not 1770 U-labels. In the presence of DNSSEC, no form of a zone file or query 1771 response that contains a U-label may be signed or the signature 1772 validated. 1774 One consequence of this for sites deploying IDNA in the presence of 1775 DNSSEC is that any special purpose proxies or forwarders used to 1776 transform user input into IDNs must be earlier in the lookup flow 1777 than DNSSEC authenticating nameservers for DNSSEC to work. 1779 8.3. Root and other DNS Server Considerations 1781 IDNs in A-label form will generally be somewhat longer than current 1782 domain names, so the bandwidth needed by the root servers is likely 1783 to go up by a small amount. Also, queries and responses for IDNs 1784 will probably be somewhat longer than typical queries historically, 1785 so EDNS0 [RFC2671] support may be more important (otherwise, queries 1786 and responses may be forced to go to TCP instead of UDP). 1788 9. Internationalization Considerations 1790 DNS labels and fully-qualified domain names provide mnemonics that 1791 assist in identifying and referring to resources on the Internet. 1792 IDNs expand the range of those mnemonics to include those based on 1793 languages and character sets other than Western European and Roman- 1794 derived ones. But domain "names" are not, in general, words in any 1795 language. The recommendations of the IETF policy on character sets 1796 and languages, BCP 18 [RFC2277] are applicable to situations in which 1797 language identification is used to provide language-specific 1798 contexts. The DNS is, by contrast, global and international and 1799 ultimately has nothing to do with languages. Adding languages (or 1800 similar context) to IDNs generally, or to DNS matching in particular, 1801 would imply context dependent matching in DNS, which would be a very 1802 significant change to the DNS protocol itself. It would also imply 1803 that users would need to identify the language associated with a 1804 particular label in order to look that label up, a decision that 1805 would be impossible in many or most cases. 1807 10. IANA Considerations 1809 This section gives an overview of registries required for IDNA. The 1810 actual definitions of the first two appear in [IDNA2008-Tables]. 1812 10.1. IDNA Character Registry 1814 The distinction among the three major categories "UNASSIGNED", 1815 "DISALLOWED", and "PROTOCOL-VALID" is made by special categories and 1816 rules that are integral elements of [IDNA2008-Tables]. Convenience 1817 in programming and validation requires a registry of characters and 1818 scripts and their categories, updated for each new version of Unicode 1819 and the characters it contains. The details of this registry are 1820 specified in [IDNA2008-Tables]. 1822 10.2. IDNA Context Registry 1824 For characters that are defined in the IDNA Character Registry list 1825 as PROTOCOL-VALID but requiring a contextual rule (i.e., the types of 1826 rule described in Section 3.1.1.1), IANA will create and maintain a 1827 list of approved contextual rules. The details for those rules 1828 appear in [IDNA2008-Tables]. 1830 10.3. IANA Repository of IDN Practices of TLDs 1832 This registry, historically described as the "IANA Language Character 1833 Set Registry" or "IANA Script Registry" (both somewhat misleading 1834 terms) is maintained by IANA at the request of ICANN. It is used to 1835 provide a central documentation repository of the IDN policies used 1836 by top level domain (TLD) registries who volunteer to contribute to 1837 it and is used in conjunction with ICANN Guidelines for IDN use. 1839 It is not an IETF-managed registry and, while the protocol changes 1840 specified here may call for some revisions to the tables, these 1841 specifications have no direct effect on that registry and no IANA 1842 action is required as a result. 1844 11. Security Considerations 1846 11.1. General Security Issues with IDNA 1848 This document in the IDNA2008 series is purely explanatory and 1849 informational and consequently introduces no new security issues. It 1850 would, of course, be a poor idea for someone to try to implement from 1851 it; such an attempt would almost certainly lead to interoperability 1852 problems and might lead to security ones. A discussion of security 1853 issues with IDNA, including some relevant history, appears in 1854 [IDNA2008-Defs]. 1856 12. Acknowledgments 1858 The editor and contributors would like to express their thanks to 1859 those who contributed significant early (pre-WG) review comments, 1860 sometimes accompanied by text, especially Mark Davis, Paul Hoffman, 1861 Simon Josefsson, and Sam Weiler. In addition, some specific ideas 1862 were incorporated from suggestions, text, or comments about sections 1863 that were unclear supplied by Vint Cerf, Frank Ellerman, Michael 1864 Everson, Asmus Freytag, Erik van der Poel, Michel Suignard, and Ken 1865 Whistler, although, as usual, they bear little or no responsibility 1866 for the conclusions the editor and contributors reached after 1867 receiving their suggestions. Thanks are also due to Vint Cerf, 1868 Debbie Garside, and Jefsey Morfin for conversations that led to 1869 considerable improvements in the content of this document. 1871 A meeting was held on 30 January 2008 to attempt to reconcile 1872 differences in perspective and terminology about this set of 1873 specifications between the design team and members of the Unicode 1874 Technical Consortium. The discussions at and subsequent to that 1875 meeting were very helpful in focusing the issues and in refining the 1876 specifications. The active participants at that meeting were (in 1877 alphabetic order as usual) Harald Alvestrand, Vint Cerf, Tina Dam, 1878 Mark Davis, Lisa Dusseault, Patrik Faltstrom (by telephone), Cary 1879 Karp, John Klensin, Warren Kumari, Lisa Moore, Erik van der Poel, 1880 Michel Suignard, and Ken Whistler. We express our thanks to Google 1881 for support of that meeting and to the participants for their 1882 contributions. 1884 Useful comments and text on the WG versions of the draft were 1885 received from many participants in the IETF "IDNABIS" WG and a number 1886 of document changes resulted from mailing list discussions made by 1887 that group. Marcos Sanz provided specific analysis and suggestions 1888 that were exceptionally helpful in refining the text, as did Vint 1889 Cerf, Mark Davis, Martin Duerst, Ken Whistler, and Andrew Sullivan. 1891 13. Contributors 1893 While the listed editor held the pen, this core of this document and 1894 the initial WG version represents the joint work and conclusions of 1895 an ad hoc design team consisting of the editor and, in alphabetic 1896 order, Harald Alvestrand, Tina Dam, Patrik Faltstrom, and Cary Karp. 1897 In addition, there were many specific contributions and helpful 1898 comments from those listed in the Acknowledgments section and others 1899 who have contributed to the development and use of the IDNA 1900 protocols. 1902 14. References 1904 14.1. Normative References 1906 [ASCII] American National Standards Institute (formerly United 1907 States of America Standards Institute), "USA Code for 1908 Information Interchange", ANSI X3.4-1968, 1968. 1910 ANSI X3.4-1968 has been replaced by newer versions with 1911 slight modifications, but the 1968 version remains 1912 definitive for the Internet. 1914 [IDNA2008-Bidi] 1915 Alvestrand, H. and C. Karp, "An updated IDNA criterion for 1916 right to left scripts", July 2008, . 1919 [IDNA2008-Defs] 1920 Klensin, J., "Internationalized Domain Names for 1921 Applications (IDNA): Definitions and Document Framework", 1922 November 2008, . 1925 [IDNA2008-Protocol] 1926 Klensin, J., "Internationalized Domain Names in 1927 Applications (IDNA): Protocol", November 2008, . 1930 [IDNA2008-Tables] 1931 Faltstrom, P., "The Unicode Code Points and IDNA", 1932 July 2008, . 1935 A version of this document is available in HTML format at 1936 http://stupid.domain.name/idnabis/ 1937 draft-ietf-idnabis-tables-02.html 1939 [RFC3490] Faltstrom, P., Hoffman, P., and A. Costello, 1940 "Internationalizing Domain Names in Applications (IDNA)", 1941 RFC 3490, March 2003. 1943 [RFC3492] Costello, A., "Punycode: A Bootstring encoding of Unicode 1944 for Internationalized Domain Names in Applications 1945 (IDNA)", RFC 3492, March 2003. 1947 [Unicode-UAX15] 1948 The Unicode Consortium, "Unicode Standard Annex #15: 1949 Unicode Normalization Forms", March 2008, 1950 . 1952 [Unicode51] 1953 The Unicode Consortium, "The Unicode Standard, Version 1954 5.1.0", 2008. 1956 defined by: The Unicode Standard, Version 5.0, Boston, MA, 1957 Addison-Wesley, 2007, ISBN 0-321-48091-0, as amended by 1958 Unicode 5.1.0 1959 (http://www.unicode.org/versions/Unicode5.1.0/). 1961 14.2. Informative References 1963 [BIG5] Institute for Information Industry of Taiwan, "Computer 1964 Chinese Glyph and Character Code Mapping Table, Technical 1965 Report C-26", 1984. 1967 There are several forms and variations and a closely- 1968 related standard, CNS 11643. See the discussion in 1969 Chapter 3 of Lunde, K., CJKV Information Processing, 1970 O'Reilly & Associates, 1999 1972 [GB18030] "Chinese National Standard GB 18030-2000: Information 1973 Technology -- Chinese ideograms coded character set for 1974 information interchange -- Extension for the basic set.", 1975 2000. 1977 [RFC0810] Feinler, E., Harrenstien, K., Su, Z., and V. White, "DoD 1978 Internet host table specification", RFC 810, March 1982. 1980 [RFC0952] Harrenstien, K., Stahl, M., and E. Feinler, "DoD Internet 1981 host table specification", RFC 952, October 1985. 1983 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 1984 STD 13, RFC 1034, November 1987. 1986 [RFC1035] Mockapetris, P., "Domain names - implementation and 1987 specification", STD 13, RFC 1035, November 1987. 1989 [RFC1123] Braden, R., "Requirements for Internet Hosts - Application 1990 and Support", STD 3, RFC 1123, October 1989. 1992 [RFC2136] Vixie, P., Thomson, S., Rekhter, Y., and J. Bound, 1993 "Dynamic Updates in the Domain Name System (DNS UPDATE)", 1994 RFC 2136, April 1997. 1996 [RFC2181] Elz, R. and R. Bush, "Clarifications to the DNS 1997 Specification", RFC 2181, July 1997. 1999 [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and 2000 Languages", BCP 18, RFC 2277, January 1998. 2002 [RFC2535] Eastlake, D., "Domain Name System Security Extensions", 2003 RFC 2535, March 1999. 2005 [RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", 2006 RFC 2671, August 1999. 2008 [RFC2673] Crawford, M., "Binary Labels in the Domain Name System", 2009 RFC 2673, August 1999. 2011 [RFC2782] Gulbrandsen, A., Vixie, P., and L. Esibov, "A DNS RR for 2012 specifying the location of services (DNS SRV)", RFC 2782, 2013 February 2000. 2015 [RFC3454] Hoffman, P. and M. Blanchet, "Preparation of 2016 Internationalized Strings ("stringprep")", RFC 3454, 2017 December 2002. 2019 [RFC3491] Hoffman, P. and M. Blanchet, "Nameprep: A Stringprep 2020 Profile for Internationalized Domain Names (IDN)", 2021 RFC 3491, March 2003. 2023 [RFC3743] Konishi, K., Huang, K., Qian, H., and Y. Ko, "Joint 2024 Engineering Team (JET) Guidelines for Internationalized 2025 Domain Names (IDN) Registration and Administration for 2026 Chinese, Japanese, and Korean", RFC 3743, April 2004. 2028 [RFC3987] Duerst, M. and M. Suignard, "Internationalized Resource 2029 Identifiers (IRIs)", RFC 3987, January 2005. 2031 [RFC4290] Klensin, J., "Suggested Practices for Registration of 2032 Internationalized Domain Names (IDN)", RFC 4290, 2033 December 2005. 2035 [RFC4690] Klensin, J., Faltstrom, P., Karp, C., and IAB, "Review and 2036 Recommendations for Internationalized Domain Names 2037 (IDNs)", RFC 4690, September 2006. 2039 [RFC4713] Lee, X., Mao, W., Chen, E., Hsu, N., and J. Klensin, 2040 "Registration and Administration Recommendations for 2041 Chinese Domain Names", RFC 4713, October 2006. 2043 [Unicode-Security] 2044 The Unicode Consortium, "Unicode Technical Standard #39: 2045 Unicode Security Mechanisms", August 2008, 2046 . 2048 [Unicode-UAX31] 2049 The Unicode Consortium, "Unicode Standard Annex #31: 2050 Unicode Identifier and Pattern Syntax", March 2008, 2051 . 2053 [Unicode-UTR36] 2054 The Unicode Consortium, "Unicode Technical Report #36: 2055 Unicode Security Considerations", July 2008, 2056 . 2058 Appendix A. Change Log 2060 [[ RFC Editor: Please remove this appendix. ]] 2062 A.1. Changes between Version -00 and Version -01 of 2063 draft-ietf-idnabis-rationale 2065 o Clarified the U-label definition to note that U-labels must 2066 contain at least one non-ASCII character. Also clarified the 2067 relationship among label types. 2069 o Rewrote the discussion of Labels in Registration (Section 7.1.2) 2070 and related text about IDNA-validity (in the "Defs" document as of 2071 -04 of this one) to narrow its focus and remove more general 2072 restrictions. Added a temporary note in line to explain the 2073 situation. 2075 o Changed the "IDNA uses Unicode" statement to focus on 2076 compatibility with IDNA2003 and avoid more general or 2077 controversial assertions. 2079 o Added a discussion of examples to Section 7.1 2081 o Made a number of other small editorial changes and corrections 2082 suggested by Mark Davis. 2084 o Added several more discussion anchors and notes and expanded or 2085 updated some existing ones. 2087 A.2. Version -02 2089 o Trimmed change log, removing information about pre-WG drafts. 2091 o Adjusted discussion of Contextual Rules to match the new location 2092 of the tables and some conceptual material. 2094 o Rewrote the material on preprocessing somewhat. 2096 o Moved the material about relationships with IDNA2003 to be part of 2097 a single section on transitions. 2099 o Removed several placeholders and made editorial changes in 2100 accordance with decisions made at IETF 72 in Dublin and not 2101 disputed on the mailing list. 2103 A.3. Version -03 2105 This special update to the Rationale document is intended to try to 2106 get the discussion of what is normative or not under control. While 2107 the IETF does not normally annotate individual sections of documents 2108 with whether they are normative or not, concerns that we don't know 2109 which is which, claims that some material is normative that would be 2110 problematic if so classified, etc., argue that we should at least be 2111 able to have a clear discussion on the subject. 2113 Two annotations have been applied to sections that might reasonably 2114 be considered normative. One annotation is based on the list of 2115 sections in Mark Davis's note of 29 September (http:// 2116 www.alvestrand.no/pipermail/idna-update/2008-September/002667.html). 2117 The other is based on an elaboration of John Klensin's response on 7 2118 October (http://www.alvestrand.no/pipermail/idna-update/2008-October/ 2119 002691.html). These should just be considered two suggestions to 2120 illuminate and, one hopes, advance the Working Group's discussions. 2122 Some additional editorial changes have been made, but they are 2123 basically trivial. In the editor's judgment, it is not possible to 2124 make significantly more progress with this document until the matter 2125 of document organization is settled. 2127 A.4. Version -04 2129 o Definitional and other normative material moved to new document 2130 (draft-ietf-idnabis-defs). Version -03 annotations removed. 2132 o Material on differences between IDNA2003 and IDNA2008 moved to an 2133 appendix in Protocol. 2135 o Material left over from the origins of this document as a 2136 preliminary proposal has been removed or rewritten. 2138 o Changes made to reflect consensus call results, including removing 2139 several placeholder notes for discussion. 2141 o Added more material, including discussion of historic scripts, to 2142 Section 3.2 on registration policies. 2144 o Added a new section (Section 7.2) to contain specific discussion 2145 of handling of characters that are interpreted differently in 2146 input to IDNA2003 and 2008. 2148 o Some material, including this section/appendix, rearranged. 2150 A.5. Version -05 2152 o Many small editorial changes, including changes to eliminate the 2153 last vestiges of what appeared to be 2119 language (upper-case 2154 MUST, SHOULD, or MAY) and small adjustments to terminology. 2156 A.6. Version -06 2158 o Removed Security Considerations material and pointed to Defs, 2159 where it now appears as of version 05. 2161 o Started changing uses of "IDNA2008" in running text to "in these 2162 specifications" or the equivalent. These documents are titled 2163 simply "IDNA"; once they are standardized, "the current version" 2164 may be a more appropriate reference than one containing a year. 2165 As discussed on the mailing list, we can and should discuss how to 2166 refer to these documents at an appropriate time (e.g., when we 2167 know when we will be finished) but, in the interim, it seems 2168 appropriate to simply start getting rid of the version-specific 2169 terminology where it can naturally be removed. 2171 o Additional discussion of mappings, etc., especially for case- 2172 sensitivity. 2174 o Clarified relationship to base DNS specifications. 2176 o Consolidated discussion of lookup of unassigned characters. 2178 o More editorial fine-tuning. 2180 A.7. Version -07 2182 o Revised terminology by adding terms: NR-LDH-label, Invalid-A-label 2183 (or False-A-label), R-LDH-label, valid IDNA-label in 2184 Section 1.3.3. 2186 o Moved the "name server considerations" material to this document 2187 from Protocol because it is non-normative and not part of the 2188 protocol itself. 2190 o To improve clarity, redid discussion of the reasons why looking up 2191 unassigned code points is prohibited. 2193 o Editorial and other non-substantive corrections to reflect earlier 2194 errors as well as new definitions and terminology. 2196 Author's Address 2198 John C Klensin 2199 1770 Massachusetts Ave, Ste 322 2200 Cambridge, MA 02140 2201 USA 2203 Phone: +1 617 245 1457 2204 Email: john+ietf@jck.com