idnits 2.17.00 (12 Aug 2021) /tmp/idnits5889/draft-ietf-idnabis-rationale-15.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Sep 2009 rather than the newer Notice from 28 Dec 2009. (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 86: '... CONTEXTUAL RULE REQUIRED . . . . . . ...' RFC 2119 keyword, line 473: '... CONTEXTUAL RULE REQUIRED, DISALLOWED ...' RFC 2119 keyword, line 499: '...3.1.2. CONTEXTUAL RULE REQUIRED...' RFC 2119 keyword, line 1633: '...nd CONTEXTUAL RULE REQUIRED categories...' RFC 2119 keyword, line 2139: '... 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 (December 15, 2009) is 4539 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ASCII' is defined on line 1886, but no explicit reference was found in the text == Unused Reference: 'Unicode-UAX15' is defined on line 1927, but no explicit reference was found in the text == Unused Reference: 'BIG5' is defined on line 1943, but no explicit reference was found in the text == Unused Reference: 'GB18030' is defined on line 1952, but no explicit reference was found in the text == Unused Reference: 'RFC0810' is defined on line 1962, but no explicit reference was found in the text == Unused Reference: 'RFC2673' is defined on line 1990, but no explicit reference was found in the text == Unused Reference: 'RFC4690' is defined on line 2020, but no explicit reference was found in the text == Unused Reference: 'Unicode-UTR36' is defined on line 2038, 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) -- No information found for draft-ietf-idnabis-mapping - is the name correct? -- Obsolete informational reference (is this intentional?): RFC 810 (Obsoleted by RFC 952) -- 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 (~~), 11 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 December 15, 2009 4 Intended status: Informational 5 Expires: June 18, 2010 7 Internationalized Domain Names for Applications (IDNA): Background, 8 Explanation, and Rationale 9 draft-ietf-idnabis-rationale-15.txt 11 Abstract 13 Several years have passed since the original protocol for 14 Internationalized Domain Names (IDNs) was completed and deployed. 15 During that time, a number of issues have arisen, including the need 16 to update the system to deal with newer versions of Unicode. Some of 17 these issues require tuning of the existing protocols and the tables 18 on which they depend. This document provides an overview of a 19 revised system and provides explanatory material for its components. 21 Status of this Memo 23 This Internet-Draft is submitted to IETF in full conformance with the 24 provisions of BCP 78 and BCP 79. 26 Internet-Drafts are working documents of the Internet Engineering 27 Task Force (IETF), its areas, and its working groups. Note that 28 other groups may also distribute working documents as Internet- 29 Drafts. 31 Internet-Drafts are draft documents valid for a maximum of six months 32 and may be updated, replaced, or obsoleted by other documents at any 33 time. It is inappropriate to use Internet-Drafts as reference 34 material or to cite them other than as "work in progress." 36 The list of current Internet-Drafts can be accessed at 37 http://www.ietf.org/ietf/1id-abstracts.txt. 39 The list of Internet-Draft Shadow Directories can be accessed at 40 http://www.ietf.org/shadow.html. 42 This Internet-Draft will expire on June 18, 2010. 44 Copyright Notice 46 Copyright (c) 2009 IETF Trust and the persons identified as the 47 document authors. All rights reserved. 49 This document is subject to BCP 78 and the IETF Trust's Legal 50 Provisions Relating to IETF Documents 51 (http://trustee.ietf.org/license-info) in effect on the date of 52 publication of this document. Please review these documents 53 carefully, as they describe your rights and restrictions with respect 54 to this document. Code Components extracted from this document must 55 include Simplified BSD License text as described in Section 4.e of 56 the Trust Legal Provisions and are provided without warranty as 57 described in the BSD License. 59 This document may contain material from IETF Documents or IETF 60 Contributions published or made publicly available before November 61 10, 2008. The person(s) controlling the copyright in some of this 62 material may not have granted the IETF Trust the right to allow 63 modifications of such material outside the IETF Standards Process. 64 Without obtaining an adequate license from the person(s) controlling 65 the copyright in such materials, this document may not be modified 66 outside the IETF Standards Process, and derivative works of it may 67 not be created outside the IETF Standards Process, except to format 68 it for publication as an RFC or to translate it into languages other 69 than English. 71 Table of Contents 73 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 74 1.1. Context and Overview . . . . . . . . . . . . . . . . . . . 4 75 1.2. Discussion Forum . . . . . . . . . . . . . . . . . . . . . 5 76 1.3. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 77 1.3.1. DNS "Name" Terminology . . . . . . . . . . . . . . . . 5 78 1.3.2. New Terminology and Restrictions . . . . . . . . . . . 6 79 1.4. Objectives . . . . . . . . . . . . . . . . . . . . . . . . 6 80 1.5. Applicability and Function of IDNA . . . . . . . . . . . . 7 81 1.6. Comprehensibility of IDNA Mechanisms and Processing . . . 8 82 2. Processing in IDNA2008 . . . . . . . . . . . . . . . . . . . . 9 83 3. Permitted Characters: An Inclusion List . . . . . . . . . . . 9 84 3.1. A Tiered Model of Permitted Characters and Labels . . . . 10 85 3.1.1. PROTOCOL-VALID . . . . . . . . . . . . . . . . . . . . 10 86 3.1.2. CONTEXTUAL RULE REQUIRED . . . . . . . . . . . . . . . 11 87 3.1.2.1. Contextual Restrictions . . . . . . . . . . . . . 11 88 3.1.2.2. Rules and Their Application . . . . . . . . . . . 12 89 3.1.3. DISALLOWED . . . . . . . . . . . . . . . . . . . . . . 12 90 3.1.4. UNASSIGNED . . . . . . . . . . . . . . . . . . . . . . 13 91 3.2. Registration Policy . . . . . . . . . . . . . . . . . . . 13 92 3.3. Layered Restrictions: Tables, Context, Registration, 93 Applications . . . . . . . . . . . . . . . . . . . . . . . 14 94 4. Issues that Constrain Possible Solutions . . . . . . . . . . . 15 95 4.1. Display and Network Order . . . . . . . . . . . . . . . . 15 96 4.2. Entry and Display in Applications . . . . . . . . . . . . 16 97 4.3. Linguistic Expectations: Ligatures, Digraphs, and 98 Alternate Character Forms . . . . . . . . . . . . . . . . 18 99 4.4. Case Mapping and Related Issues . . . . . . . . . . . . . 20 100 4.5. Right to Left Text . . . . . . . . . . . . . . . . . . . . 21 101 5. IDNs and the Robustness Principle . . . . . . . . . . . . . . 21 102 6. Front-end and User Interface Processing for Lookup . . . . . . 22 103 7. Migration from IDNA2003 and Unicode Version Synchronization . 24 104 7.1. Design Criteria . . . . . . . . . . . . . . . . . . . . . 24 105 7.1.1. Summary and Discussion of IDNA Validity Criteria . . . 24 106 7.1.2. Labels in Registration . . . . . . . . . . . . . . . . 25 107 7.1.3. Labels in Lookup . . . . . . . . . . . . . . . . . . . 26 108 7.2. Changes in Character Interpretations . . . . . . . . . . . 28 109 7.2.1. Character Changes: Eszett and Final Sigma . . . . . . 28 110 7.2.2. Character Changes: Zero-Width Joiner and Non-Joiner . 28 111 7.2.3. Character Changes and the Need for Transition . . . . 28 112 7.2.4. Transition Strategies . . . . . . . . . . . . . . . . 29 113 7.3. Elimination of Character Mapping . . . . . . . . . . . . . 30 114 7.4. The Question of Prefix Changes . . . . . . . . . . . . . . 30 115 7.4.1. Conditions Requiring a Prefix Change . . . . . . . . . 31 116 7.4.2. Conditions Not Requiring a Prefix Change . . . . . . . 31 117 7.4.3. Implications of Prefix Changes . . . . . . . . . . . . 32 118 7.5. Stringprep Changes and Compatibility . . . . . . . . . . . 32 119 7.6. The Symbol Question . . . . . . . . . . . . . . . . . . . 33 120 7.7. Migration Between Unicode Versions: Unassigned Code 121 Points . . . . . . . . . . . . . . . . . . . . . . . . . . 34 122 7.8. Other Compatibility Issues . . . . . . . . . . . . . . . . 36 123 8. Name Server Considerations . . . . . . . . . . . . . . . . . . 36 124 8.1. Processing Non-ASCII Strings . . . . . . . . . . . . . . . 36 125 8.2. Root and other DNS Server Considerations . . . . . . . . . 37 126 9. Internationalization Considerations . . . . . . . . . . . . . 37 127 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 38 128 10.1. IDNA Character Registry . . . . . . . . . . . . . . . . . 38 129 10.2. IDNA Context Registry . . . . . . . . . . . . . . . . . . 38 130 10.3. IANA Repository of IDN Practices of TLDs . . . . . . . . . 38 131 11. Security Considerations . . . . . . . . . . . . . . . . . . . 38 132 11.1. General Security Issues with IDNA . . . . . . . . . . . . 38 133 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 39 134 13. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 39 135 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 40 136 14.1. Normative References . . . . . . . . . . . . . . . . . . . 40 137 14.2. Informative References . . . . . . . . . . . . . . . . . . 41 138 Appendix A. Change Log . . . . . . . . . . . . . . . . . . . . . 43 139 A.1. Changes between Version -00 and Version -01 of 140 draft-ietf-idnabis-rationale . . . . . . . . . . . . . . . 43 141 A.2. Version -02 . . . . . . . . . . . . . . . . . . . . . . . 44 142 A.3. Version -03 . . . . . . . . . . . . . . . . . . . . . . . 44 143 A.4. Version -04 . . . . . . . . . . . . . . . . . . . . . . . 45 144 A.5. Version -05 . . . . . . . . . . . . . . . . . . . . . . . 45 145 A.6. Version -06 . . . . . . . . . . . . . . . . . . . . . . . 45 146 A.7. Version -07 . . . . . . . . . . . . . . . . . . . . . . . 46 147 A.8. Version -08 . . . . . . . . . . . . . . . . . . . . . . . 46 148 A.9. Version -09 . . . . . . . . . . . . . . . . . . . . . . . 46 149 A.10. Version -10 . . . . . . . . . . . . . . . . . . . . . . . 47 150 A.11. Version -11 . . . . . . . . . . . . . . . . . . . . . . . 47 151 A.12. Version -12 . . . . . . . . . . . . . . . . . . . . . . . 48 152 A.13. Version -13 . . . . . . . . . . . . . . . . . . . . . . . 48 153 A.14. Version -14 . . . . . . . . . . . . . . . . . . . . . . . 48 154 A.15. Version -15 . . . . . . . . . . . . . . . . . . . . . . . 49 155 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 49 157 1. Introduction 159 1.1. Context and Overview 161 Internationalized Domain Names in Applications (IDNA) is a collection 162 of standards that allow client applications to convert some Unicode 163 mnemonics to an ASCII-compatible encoding form ("ACE") which is a 164 valid DNS label containing only letters, digits, and hyphens. The 165 specific form of ACE label used by IDNA is called an "A-label". A 166 client can look up an exact A-label in the existing DNS, so A-labels 167 do not require any extensions to DNS, upgrades of DNS servers or 168 updates to low-level client libraries. An A-label is recognizable 169 from the prefix "xn--" before the characters produced by the Punycode 170 algorithm [RFC3492], thus a user application can identify an A-label 171 and convert it into Unicode (or some local coded character set) for 172 display. 174 On the registry side, IDNA allows a registry to offer 175 Internationalized Domain Names (IDNs) for registration as A-labels. 176 A registry may offer any subset of valid IDNs, and may apply any 177 restrictions or bundling (grouping of similar labels together in one 178 registration) appropriate for the context of that registry. 179 Registration of labels is sometimes discussed separately from lookup, 180 and is subject to a few specific requirements that do not apply to 181 lookup. 183 DNS clients and registries are subject to some differences in 184 requirements for handling IDNs. In particular, registries are urged 185 to register only exact, valid A-labels, while clients might do some 186 mapping to get from otherwise-invalid user input to a valid A-label. 188 The first version of IDNA was published in 2003 and is referred to 189 here as IDNA2003 to contrast it with the current version, which is 190 known as IDNA2008 (after the year in which IETF work started on it). 191 IDNA2003 consists of four documents: the IDNA base specification 192 [RFC3490], Nameprep [RFC3491], Punycode [RFC3492], and Stringprep 193 [RFC3454]. The current set of documents, IDNA2008, are not dependent 194 on any of the IDNA2003 specifications other than the one for Punycode 195 encoding. References to "these specifications" or "these documents" 196 are to the entire IDNA2008 set listed in [IDNA2008-Defs]. The 197 characters that are valid in A-labels are identified from rules 198 listed in the Tables document [IDNA2008-Tables], but validity can be 199 derived from the Unicode properties of those characters with a very 200 few exceptions. 202 Traditionally, DNS labels are matched case-insensitively 203 [RFC1034][RFC1035]. That convention was preserved in IDNA2003 by a 204 case-folding operation that generally maps capital letters into 205 lower-case ones. However, if case rules are enforced from one 206 language, another language sometimes loses the ability to treat two 207 characters separately. Case-insensitivity is treated slightly 208 differently in IDNA2008. 210 IDNA2003 used Unicode version 3.2 only. In order to keep up with new 211 characters added in new versions of UNICODE, IDNA2008 decouples its 212 rules from any particular version of UNICODE. Instead, the 213 attributes of new characters in Unicode, supplemented by a small 214 number of exception cases, determine how and whether the characters 215 can be used in IDNA labels. 217 This document provides informational context for IDNA2008, including 218 terminology, background, and policy discussions. 220 1.2. Discussion Forum 222 [[ RFC Editor: please remove this section. ]] 224 IDNA2008 is being discussed in the IETF "idnabis" Working Group and 225 on the mailing list idna-update@alvestrand.no 227 1.3. Terminology 229 Terminology for IDNA2008 appears in [IDNA2008-Defs]. That document 230 also contains a roadmap to the IDNA2008 document collection. No 231 attempt should be made to understand this document without the 232 definitions and concepts that appear there. 234 1.3.1. DNS "Name" Terminology 236 In the context of IDNs, the DNS term "name" has introduced some 237 confusion as people speak of DNS labels in terms of the words or 238 phrases of various natural languages. Historically, many of the 239 "names" in the DNS have been mnemonics to identify some particular 240 concept, object, or organization. They are typically rooted in some 241 language because most people think in language-based ways. But, 242 because they are mnemonics, they need not obey the orthographic 243 conventions of any language: it is not a requirement that it be 244 possible for them to be "words". 246 This distinction is important because the reasonable goal of an IDN 247 effort is not to be able to write the great Klingon (or language of 248 one's choice) novel in DNS labels but to be able to form a usefully 249 broad range of mnemonics in ways that are as natural as possible in a 250 very broad range of scripts. 252 1.3.2. New Terminology and Restrictions 254 These documents introduce new terminology, and precise definitions 255 (in [IDNA2008-Defs]), for the terms "U-label", "A-Label", LDH-label 256 (to which all valid pre-IDNA host names conformed), Reserved-LDH- 257 label (R-LDH-label), XN-label, Fake-A-Label, and Non-Reserved-LDH- 258 label (NR-LDH-label). 260 In addition, the term "putative label" has been adopted to refer to a 261 label that may appear to meet certain definitional constraints but 262 has not yet been sufficiently tested for validity. 264 These definitions are also illustrated in Figure 1 of the Definitions 265 Document [IDNA2008-Defs]. R-LDH-labels contain "--" in the third and 266 fourth character from the beginning of the label. In IDNA-aware 267 applications, only a subset of these reserved labels is permitted to 268 be used, namely the A-label subset. A-labels are a subset of the 269 R-LDH-labels that begin with the case-insensitive string "xn--". 270 Labels that bear this prefix but which are not otherwise valid fall 271 into the "Fake A-label" category. The non-reserved labels (NR-LDH- 272 labels) are implicitly valid since they do not bear any resemblance 273 to the labels specified by IDNA. 275 The creation of the Reserved-LDH category is required for three 276 reasons: 278 o to prevent confusion with pre-IDNA coding forms; 280 o to permit future extensions that would require changing the 281 prefix, no matter how unlikely those might be (see Section 7.4); 282 and 284 o to reduce the opportunities for attacks via the Punycode encoding 285 algorithm itself. 287 As with other documents in the IDNA2008 set, this document uses the 288 term "registry" to describe any zone in the DNS. That term, and the 289 terms "zone" or "zone administration", are interchangeable. 291 1.4. Objectives 293 These are the main objectives in revising IDNA. 295 o Use a more recent version of Unicode, and allow IDNA to be 296 independent of Unicode versions, so that IDNA2008 need not be 297 updated for implementations to adopt codepoints from new Unicode 298 versions. 300 o Fix a very small number of code-point categorizations that have 301 turned out to cause problems in the communities that use those 302 code-points. 304 o Reduce the dependency on mapping, in order that the pre-mapped 305 forms (which are not valid IDNA labels) tend to appear less often 306 in various contexts, in favor of valid A-labels. 308 o Fix some details in the bidirectional codepoint handling 309 algorithms. 311 1.5. Applicability and Function of IDNA 313 The IDNA specification solves the problem of extending the repertoire 314 of characters that can be used in domain names to include a large 315 subset of the Unicode repertoire. 317 IDNA does not extend DNS. Instead, the applications (and, by 318 implication, the users) continue to see an exact-match lookup 319 service. Either there is a single exactly-matching name (subject to 320 the base DNS requirement of case-insensitive ASCII matching) or there 321 is no match. This model has served the existing applications well, 322 but it requires, with or without internationalized domain names, that 323 users know the exact spelling of the domain names that are to be 324 typed into applications such as web browsers and mail user agents. 325 The introduction of the larger repertoire of characters potentially 326 makes the set of misspellings larger, especially given that in some 327 cases the same appearance, for example on a business card, might 328 visually match several Unicode code points or several sequences of 329 code points. 331 The IDNA standard does not require any applications to conform to it, 332 nor does it retroactively change those applications. An application 333 can elect to use IDNA in order to support IDN while maintaining 334 interoperability with existing infrastructure. If an application 335 wants to use non-ASCII characters in public DNS domain names, IDNA is 336 the only currently-defined option. Adding IDNA support to an 337 existing application entails changes to the application only, and 338 leaves room for flexibility in front-end processing and more 339 specifically in the user interface (see Section 6). 341 A great deal of the discussion of IDN solutions has focused on 342 transition issues and how IDNs will work in a world where not all of 343 the components have been updated. Proposals that were not chosen by 344 the original IDN Working Group would have depended on updating of 345 user applications, DNS resolvers, and DNS servers in order for a user 346 to apply an internationalized domain name in any form or coding 347 acceptable under that method. While processing must be performed 348 prior to or after access to the DNS, IDNA requires no changes to the 349 DNS protocol, any DNS servers, or the resolvers on users' computers. 351 IDNA allows the graceful introduction of IDNs not only by avoiding 352 upgrades to existing infrastructure (such as DNS servers and mail 353 transport agents), but also by allowing some limited use of IDNs in 354 applications by using the ASCII-encoded representation of the labels 355 containing non-ASCII characters. While such names are user- 356 unfriendly to read and type, and hence not optimal for user input, 357 they can be used as a last resort to allow rudimentary IDN usage. 358 For example, they might be the best choice for display if it were 359 known that relevant fonts were not available on the user's computer. 360 In order to allow user-friendly input and output of the IDNs and 361 acceptance of some characters as equivalent to those to be processed 362 according to the protocol, the applications need to be modified to 363 conform to this specification. 365 This version of IDNA uses the Unicode character repertoire, for 366 continuity with the original version of IDNA. 368 1.6. Comprehensibility of IDNA Mechanisms and Processing 370 One goal of IDNA2008, which is aided by the main goal of reducing the 371 dependency on mapping, is to improve the general understanding of how 372 IDNA works and what characters are permitted and what happens to 373 them. Comprehensibility and predictability to users and registrants 374 are important design goals for this effort. End-user applications 375 have an important role to play in increasing this comprehensibility. 377 Any system that tries to handle international characters encounters 378 some common problems. For example, a UI cannot display a character 379 if no font for that character is available. In some cases, 380 internationalization enables effective localization while maintaining 381 some global uniformity but losing some universality. 383 It is difficult to even make suggestions for end-user applications to 384 cope when characters and fonts are not available. Because display 385 functions are rarely controlled by the types of applications that 386 would call upon IDNA, such suggestions will rarely be very effective. 388 Converting between local character sets and normalized Unicode, if 389 needed, is part of this set of user agent issues. This conversion 390 introduces complexity in a system that is not Unicode-native. If a 391 label is converted to a local character set that does not have all 392 the needed characters, or that uses different character-coding 393 principles, the user agent may have to add special logic to avoid or 394 reduce loss of information. 396 The major difficulty may lie in accurately identifying the incoming 397 character set and applying the correct conversion routine. Even more 398 difficult, the local character coding system could be based on 399 conceptually different assumptions than those used by Unicode (e.g., 400 choice of font encodings used for publications in some Indic 401 scripts). Those differences may not easily yield unambiguous 402 conversions or interpretations even if each coding system is 403 internally consistent and adequate to represent the local language 404 and script. 406 IDNA2008 shifts responsibility for character mapping and other 407 adjustments from the protocol (where it was located in IDNA2003) to 408 pre-processing before invoking IDNA itself. The intent is that this 409 change will lead to greater usage of fully-valid A-Labels or U-labels 410 in display, transit and storage, which should aid comprehensibility 411 and predictability. A careful look at pre-processing raises issues 412 about what that pre-processing should do and at what point pre- 413 processing becomes harmful, how universally consistent pre-processing 414 algorithms can be, and how to be compatible with labels prepared in a 415 IDNA2003 context. Those issues are discussed in Section 6 and in the 416 separate document [IDNA2008-Mapping]. 418 2. Processing in IDNA2008 420 These specifications separate Domain Name Registration and Lookup in 421 the protocol specification. Although most steps in the two processes 422 are similar, the separation reflects current practice in which per- 423 registry (DNS zone) restrictions and special processing are applied 424 at registration time but not during lookup. Another significant 425 benefit is that separation facilitates incremental addition of 426 permitted character groups to avoid freezing on one particular 427 version of Unicode. 429 The actual registration and lookup protocols for IDNA2008 are 430 specified in [IDNA2008-Protocol]. 432 3. Permitted Characters: An Inclusion List 434 IDNA2008 adopts the inclusion model. A code-point is assumed to be 435 invalid for IDN use unless it is included as part of a Unicode 436 property-based rule or, in rare cases, included individually by an 437 exception. When an implementation moves to a new version of Unicode, 438 the rules may indicate new valid code-points. 440 This section provides an overview of the model used to establish the 441 algorithm and character lists of [IDNA2008-Tables] and describes the 442 names and applicability of the categories used there. Note that the 443 inclusion of a character in the first category group (Section 3.1.1) 444 does not imply that it can be used indiscriminately; some characters 445 are associated with contextual rules that must be applied as well. 447 The information given in this section is provided to make the rules, 448 tables, and protocol easier to understand. The normative generating 449 rules that correspond to this informal discussion appear in 450 [IDNA2008-Tables] and the rules that actually determine what labels 451 can be registered or looked up are in [IDNA2008-Protocol]. 453 3.1. A Tiered Model of Permitted Characters and Labels 455 Moving to an inclusion model involves a new specification for the 456 list of characters that are permitted in IDNs. In IDNA2003, 457 character validity is independent of context and fixed forever (or 458 until the standard is replaced). However, globally context- 459 independent rules have proved to be impractical because some 460 characters, especially those that are called "Join_Controls" in 461 Unicode, are needed to make reasonable use of some scripts but have 462 no visible effect in others. IDNA2003 prohibited those types of 463 characters entirely by discarding them. We now have a consensus that 464 under some conditions, these "joiner" characters are legitimately 465 needed to allow useful mnemonics for some languages and scripts. In 466 general, context-dependent rules help deal with characters (generally 467 characters that would otherwise be prohibited entirely) that are used 468 differently or perceived differently across different scripts, and 469 allow the standard to be applied more appropriately in cases where a 470 string is not universally handled the same way. 472 IDNA2008 divides all possible Unicode code-points into four 473 categories: PROTOCOL-VALID, CONTEXTUAL RULE REQUIRED, DISALLOWED and 474 UNASSIGNED. 476 3.1.1. PROTOCOL-VALID 478 Characters identified as "PROTOCOL-VALID" (often abbreviated 479 "PVALID") are permitted in IDNs. Their use may be restricted by 480 rules about the context in which they appear or by other rules that 481 apply to the entire label in which they are to be embedded. For 482 example, any label that contains a character in this category that 483 has a "right-to-left" property must be used in context with the 484 "Bidi" rules (see [IDNA2008-Bidi]). 486 The term "PROTOCOL-VALID" is used to stress the fact that the 487 presence of a character in this category does not imply that a given 488 registry need accept registrations containing any of the characters 489 in the category. Registries are still expected to apply judgment 490 about labels they will accept and to maintain rules consistent with 491 those judgments (see [IDNA2008-Protocol] and Section 3.3). 493 Characters that are placed in the "PROTOCOL-VALID" category are 494 expected to never be removed from it or reclassified. While 495 theoretically characters could be removed from Unicode, such removal 496 would be inconsistent with the Unicode stability principles (see 497 [Unicode51], Appendix F) and hence should never occur. 499 3.1.2. CONTEXTUAL RULE REQUIRED 501 Some characters may be unsuitable for general use in IDNs but 502 necessary for the plausible support of some scripts. The two most 503 commonly-cited examples are the zero-width joiner and non-joiner 504 characters (ZWJ, U+200D and ZWNJ, U+200C) but other characters may 505 require special treatment because they would otherwise be DISALLOWED 506 (typically because Unicode considers them punctuation or special 507 symbols) but need to be permitted in limited contexts. Other 508 characters are given this special treatment because they pose 509 exceptional danger of being used to produce misleading labels or to 510 cause unacceptable ambiguity in label matching and interpretation. 512 3.1.2.1. Contextual Restrictions 514 Characters with contextual restrictions are identified as "CONTEXTUAL 515 RULE REQUIRED" and associated with a rule. The rule defines whether 516 the character is valid in a particular string, and also whether the 517 rule itself is to be applied on lookup as well as registration. 519 A distinction is made between characters that indicate or prohibit 520 joining and ones similar to them (known as "CONTEXT-JOINER" or 521 "CONTEXTJ") and other characters requiring contextual treatment 522 ("CONTEXT-OTHER" or "CONTEXTO"). Only the former require full 523 testing at lookup time. 525 It is important to note that these contextual rules cannot prevent 526 all uses of the relevant characters that might be confusing or 527 problematic. What they are expected to do is to confine 528 applicability of the characters to scripts (and narrower contexts) 529 where zone administrators are knowledgeable enough about the use of 530 those characters to be prepared to deal with them appropriately. 532 For example, a registry dealing with an Indic script that requires 533 ZWJ and/or ZWNJ as part of the writing system is expected to 534 understand where the characters have visible effect and where they do 535 not and to make registration rules accordingly. By contrast, a 536 registry dealing primarily with Latin or Cyrillic script might not be 537 actively aware that the characters exist, much less about the 538 consequences of embedding them in labels drawn from those scripts and 539 therefore should avoid accepting registrations containing those 540 characters, at least in Latin or Cyrillic-script labels. 542 3.1.2.2. Rules and Their Application 544 Rules have descriptions such as "Must follow a character from Script 545 XYZ", "Must occur only if the entire label is in Script ABC", or 546 "Must occur only if the previous and subsequent characters have the 547 DFG property". The actual rules may be DEFINED or NULL. If present, 548 they may have values of "True" (character may be used in any position 549 in any label), "False" (character may not be used in any label), or 550 may be a set of procedural rules that specify the context in which 551 the character is permitted. 553 Because it is easier to identify these characters than to know that 554 they are actually needed in IDNs or how to establish exactly the 555 right rules for each one, a rule may have a null value in a given 556 version of the tables. Characters associated with null rules are not 557 permitted to appear in putative labels for either registration or 558 lookup. Of course, a later version of the tables might contain a 559 non-null rule. 561 The actual rules and their descriptions are in Sections 2 and 3 of 562 [IDNA2008-Tables]. That document also specifies the creation of a 563 registry for future rules. 565 3.1.3. DISALLOWED 567 Some characters are inappropriate for use in IDNs and are thus 568 excluded for both registration and lookup (i.e., IDNA-conforming 569 applications performing name lookup should verify that these 570 characters are absent; if they are present, the label strings should 571 be rejected rather than converted to A-labels and looked up. Some of 572 these characters are problematic for use in IDNs (such as the 573 FRACTION SLASH character, U+2044), while some of them (such as the 574 various HEART symbols, e.g., U+2665, U+2661, and U+2765, see 575 Section 7.6) simply fall outside the conventions for typical 576 identifiers (basically letters and numbers). 578 Of course, this category would include code points that had been 579 removed entirely from Unicode should such removals ever occur. 581 Characters that are placed in the "DISALLOWED" category are expected 582 to never be removed from it or reclassified. If a character is 583 classified as "DISALLOWED" in error and the error is sufficiently 584 problematic, the only recourse would be either to introduce a new 585 code point into Unicode and classify it as "PROTOCOL-VALID" or for 586 the IETF to accept the considerable costs of an incompatible change 587 and replace the relevant RFC with one containing appropriate 588 exceptions. 590 There is provision for exception cases but, in general, characters 591 are placed into "DISALLOWED" if they fall into one or more of the 592 following groups: 594 o The character is a compatibility equivalent for another character. 595 In slightly more precise Unicode terms, application of 596 normalization method NFKC to the character yields some other 597 character. 599 o The character is an upper-case form or some other form that is 600 mapped to another character by Unicode casefolding. 602 o The character is a symbol or punctuation form or, more generally, 603 something that is not a letter, digit, or a mark that is used to 604 form a letter or digit. 606 3.1.4. UNASSIGNED 608 For convenience in processing and table-building, code points that do 609 not have assigned values in a given version of Unicode are treated as 610 belonging to a special UNASSIGNED category. Such code points are 611 prohibited in labels to be registered or looked up. The category 612 differs from DISALLOWED in that code points are moved out of it by 613 the simple expedient of being assigned in a later version of Unicode 614 (at which point, they are classified into one of the other categories 615 as appropriate). 617 The rationale for restricting the processing of UNASSIGNED characters 618 is simply that the properties of such code points cannot be 619 completely known until actual characters are assigned to them. For 620 example, assume that an UNASSIGNED code point were included in a 621 label to be looked up. Assume that the code point was later assigned 622 to a character that required some set of contextual rules. With that 623 combination, un-updated instances of IDNA-aware software might permit 624 lookup of labels containing the previously-unassigned characters 625 while updated versions of the software might restrict use of the same 626 label in lookup, depending on the contextual rules. It should be 627 clear that under no circumstance should an UNASSIGNED character be 628 permitted in a label to be registered as part of a domain name. 630 3.2. Registration Policy 632 While these recommendations cannot and should not define registry 633 policies, registries should develop and apply additional restrictions 634 as needed to reduce confusion and other problems. For example, it is 635 generally believed that labels containing characters from more than 636 one script are a bad practice although there may be some important 637 exceptions to that principle. Some registries may choose to restrict 638 registrations to characters drawn from a very small number of 639 scripts. For many scripts, the use of variant techniques such as 640 those as described in RFC 3743 [RFC3743] and RFC 4290 [RFC4290], and 641 illustrated for Chinese by the tables described in RFC 4713 [RFC4713] 642 may be helpful in reducing problems that might be perceived by users. 644 In general, users will benefit if registries only permit characters 645 from scripts that are well-understood by the registry or its 646 advisers. If a registry decides to reduce opportunities for 647 confusion by constructing policies that disallow characters used in 648 historic writing systems or characters whose use is restricted to 649 specialized, highly technical contexts, some relevant information may 650 be found in Section 2.4 "Specific Character Adjustments", Table 4 651 "Candidate Characters for Exclusion from Identifiers" of 652 [Unicode-UAX31] and Section 3.1. "General Security Profile for 653 Identifiers" in [Unicode-Security]. 655 The requirement (in Section 4.1 of [IDNA2008-Protocol]) that 656 registration procedures use only U-labels and/or A-labels is intended 657 to ensure that registrants are fully aware of exactly what is being 658 registered as well as encouraging use of those canonical forms. That 659 provision should not be interpreted as requiring that registrants 660 need to provide characters in a particular code sequence. Registrant 661 input conventions and management are part of registrant-registrar 662 interactions and relationships between registries and registrars and 663 are outside the scope of these standards. 665 It is worth stressing that these principles of policy development and 666 application apply at all levels of the DNS, not only, e.g., TLD or 667 SLD registrations. Even a trivial, "anything is permitted that is 668 valid under the protocol" policy is helpful in that it helps users 669 and application developers know what to expect. 671 3.3. Layered Restrictions: Tables, Context, Registration, Applications 673 The character rules in IDNA2008 are based on the realization that 674 there is no single magic bullet for any of the security, 675 confusability, or other issues associated with IDNs. Instead, the 676 specifications define a variety of approaches. The character tables 677 are the first mechanism, protocol rules about how those characters 678 are applied or restricted in context are the second, and those two in 679 combination constitute the limits of what can be done in the 680 protocol. As discussed in the previous section (Section 3.2), 681 registries are expected to restrict what they permit to be 682 registered, devising and using rules that are designed to optimize 683 the balance between confusion and risk on the one hand and maximum 684 expressiveness in mnemonics on the other. 686 In addition, there is an important role for user agents in warning 687 against label forms that appear problematic given their knowledge of 688 local contexts and conventions. Of course, no approach based on 689 naming or identifiers alone can protect against all threats. 691 4. Issues that Constrain Possible Solutions 693 4.1. Display and Network Order 695 Domain names are always transmitted in network order (the order in 696 which the code points are sent in protocols), but may have a 697 different display order (the order in which the code points are 698 displayed on a screen or paper). When a domain name contains 699 characters that are normally written right to left, display order may 700 be affected although network order is not. It gets even more 701 complicated if left to right and right to left labels are adjacent to 702 each other within a domain name. The decision about the display 703 order is ultimately under the control of user agents --including Web 704 browsers, mail clients, hosted Web applications and many more -- 705 which may be highly localized. Should a domain name abc.def, in 706 which both labels are represented in scripts that are written right 707 to left, be displayed as fed.cba or cba.fed? Applications that are 708 in deployment today are already diverse, and one can find examples of 709 either choice. 711 The picture changes once again when an IDN appears in a 712 Internationalized Resource Identifier (IRI) [RFC3987]. An IRI or 713 Internationalized Email address contains elements other than the 714 domain name. For example, IRIs contain protocol identifiers and 715 field delimiter syntax such as "http://" or "mailto:" while email 716 addresses contain the "@" to separate local parts from domain names. 717 An IRI in network order begins with "http://" followed by domain 718 labels in network order, thus "http://abc.def". 720 User agents are not required to display and allow input of IRIs 721 directly but often do so. Implementors have to choose whether the 722 overall direction of these strings will always be left to right (or 723 right to left) for an IRI or email address. The natural order for a 724 user typing a domain name on a right to left system is fed.cba. 725 Should the R2L user agent reverse the entire domain name each time a 726 domain name is typed? Does this change if the user types "http://" 727 right before typing a domain name, thus implying that the user is 728 beginning at the beginning of the network order IRI? Experience in 729 the 1980s and 1990s with mixing systems in which domain name labels 730 were read in network order (left to right) and those in which those 731 labels were read right to left would predict a great deal of 732 confusion. 734 If each implementation of each application makes its own decisions on 735 these issues, users will develop heuristics that will sometimes fail 736 when switching applications. However, while some display order 737 conventions, voluntarily adopted, would be desirable to reduce 738 confusion, such suggestions are beyond the scope of these 739 specifications. 741 4.2. Entry and Display in Applications 743 Applications can accept and display domain names using any character 744 set or character coding system. The IDNA protocol does not 745 necessarily affect the interface between users and applications. An 746 IDNA-aware application can accept and display internationalized 747 domain names in two formats: the internationalized character set(s) 748 supported by the application (i.e., an appropriate local 749 representation of a U-label), and as an A-label. Applications may 750 allow the display of A-labels, but are encouraged to not do so except 751 as an interface for special purposes, possibly for debugging, or to 752 cope with display limitations. In general, they should allow, but 753 not encourage, user input of A-labels. A-labels are opaque, ugly, 754 and malicious variations on them are not easily detected by users. 755 Where possible, they should thus only be exposed when they are 756 absolutely needed. Because IDN labels can be rendered either as 757 A-labels or U-labels, the application may reasonably have an option 758 for the user to select the preferred method of display. Rendering 759 the U-label should normally be the default. 761 Domain names are often stored and transported in many places. For 762 example, they are part of documents such as mail messages and web 763 pages. They are transported in many parts of many protocols, such as 764 both the control commands of SMTP and associated message body parts, 765 and in the headers and the body content in HTTP. It is important to 766 remember that domain names appear both in domain name slots and in 767 the content that is passed over protocols and it would be helpful if 768 protocols explicitly define what their domain name slots are. 770 In protocols and document formats that define how to handle 771 specification or negotiation of charsets, labels can be encoded in 772 any charset allowed by the protocol or document format. If a 773 protocol or document format only allows one charset, the labels must 774 be given in that charset. Of course, not all charsets can properly 775 represent all labels. If a U-label cannot be displayed in its 776 entirety, the only choice (without loss of information) may be to 777 display the A-label. 779 Where a protocol or document format allows IDNs, labels should be in 780 whatever character encoding and escape mechanism the protocol or 781 document format uses at that place. This provision is intended to 782 prevent situations in which, e.g., UTF-8 domain names appear embedded 783 in text that is otherwise in some other character coding. 785 All protocols that use domain name slots (See Section 2.3.1.6 in 786 [IDNA2008-Defs]) already have the capacity for handling domain names 787 in the ASCII charset. Thus, A-labels can inherently be handled by 788 those protocols. 790 These documents do not specify required mappings between one 791 character or code point and others. An extended discussion of 792 mapping issues occurs in Section 6 and specific recommendations 793 appear in [IDNA2008-Mapping]. In general, IDNA2008 prohibits 794 characters that would be mapped to others by normalization or other 795 rules. As examples, while mathematical characters based on Latin 796 ones are accepted as input to IDNA2003, they are prohibited in 797 IDNA2008. Similarly, upper-case characters, double-width characters, 798 and other variations are prohibited as IDNA input although mapping 799 them as needed in user interfaces is strongly encouraged. 801 Since the rules in [IDNA2008-Tables] have the effect that only 802 strings that are not transformed by NFKC are valid, if an application 803 chooses to perform NFKC normalization before lookup, that operation 804 is safe since this will never make the application unable to look up 805 any valid string. However, as discussed above, the application 806 cannot guarantee that any other application will perform that 807 mapping, so it should be used only with caution and for informed 808 users. 810 In many cases these prohibitions should have no effect on what the 811 user can type as input to the lookup process. It is perfectly 812 reasonable for systems that support user interfaces to perform some 813 character mapping that is appropriate to the local environment. This 814 would normally be done prior to actual invocation of IDNA. At least 815 conceptually, the mapping would be part of the Unicode conversions 816 discussed above and in [IDNA2008-Protocol]. However, those changes 817 will be local ones only -- local to environments in which users will 818 clearly understand that the character forms are equivalent. For use 819 in interchange among systems, it appears to be much more important 820 that U-labels and A-labels can be mapped back and forth without loss 821 of information. 823 One specific, and very important, instance of this strategy arises 824 with case-folding. In the ASCII-only DNS, names are looked up and 825 matched in a case-independent way, but no actual case-folding occurs. 826 Names can be placed in the DNS in either upper or lower case form (or 827 any mixture of them) and that form is preserved, returned in queries, 828 and so on. IDNA2003 approximated that behavior for non-ASCII strings 829 by performing case-folding at registration time (resulting in only 830 lower-case IDNs in the DNS) and when names were looked up. 832 As suggested earlier in this section, it appears to be desirable to 833 do as little character mapping as possible as long as Unicode works 834 correctly (e.g., NFC mapping to resolve different codings for the 835 same character is still necessary although the specifications require 836 that it be performed prior to invoking the protocol) in order to make 837 the mapping between A-labels and U-labels idempotent. Case-mapping 838 is not an exception to this principle. If only lower case characters 839 can be registered in the DNS (i.e., be present in a U-label), then 840 IDNA2008 should prohibit upper-case characters as input even though 841 user interfaces to applications should probably map those characters. 842 Some other considerations reinforce this conclusion. For example, in 843 ASCII case-mapping for individual characters, uppercase(character) 844 must be equal to uppercase(lowercase(character)). That may not be 845 true with IDNs. In some scripts that use case distinctions, there 846 are a few characters that do not have counterparts in one case or the 847 other. The relationship between upper case and lower case may even 848 be language-dependent, with different languages (or even the same 849 language in different areas) expecting different mappings. User 850 agents can meet the expectations of users who are accustomed to the 851 case-insensitive DNS environment by performing case folding prior to 852 IDNA processing, but the IDNA procedures themselves should neither 853 require such mapping nor expect them when they are not natural to the 854 localized environment. 856 4.3. Linguistic Expectations: Ligatures, Digraphs, and Alternate 857 Character Forms 859 Users have expectations about character matching or equivalence that 860 are based on their own languages and the orthography of those 861 languages. These expectations may not always be met in a global 862 system, especially if multiple languages are written using the same 863 script but using different conventions. Some examples: 865 o A Norwegian user might expect a label with the ae-ligature to be 866 treated as the same label as one using the Swedish spelling with 867 a-diaeresis even though applying that mapping to English would be 868 astonishing to users. 870 o A user in German might expect a label with an o-umlaut and a label 871 that had "oe" substituted, but was otherwise the same, treated as 872 equivalent even though that substitution would be a clear error in 873 Swedish. 875 o A Chinese user might expect automatic matching of Simplified and 876 Traditional Chinese characters, but applying that matching for 877 Korean or Japanese text would create considerable confusion. 879 o An English user might expect "theater" and "theatre" to match. 881 A number of languages use alphabetic scripts in which single phonemes 882 are written using two characters, termed a "digraph", for example, 883 the "ph" in "pharmacy" and "telephone". (Such characters can also 884 appear consecutively without forming a digraph, as in "tophat".) 885 Certain digraphs may be indicated typographically by setting the two 886 characters closer together than they would be if used consecutively 887 to represent different phonemes. Some digraphs are fully joined as 888 ligatures. For example, the word "encyclopaedia" is sometimes set 889 with a U+00E6 LATIN SMALL LIGATURE AE. When ligature and digraph 890 forms have the same interpretation across all languages that use a 891 given script, application of Unicode normalization generally resolves 892 the differences and causes them to match. When they have different 893 interpretations, matching must utilize other methods, presumably 894 chosen at the registry level, or users must be educated to understand 895 that matching will not occur. 897 The nature of the problem can be illustrated by many words in the 898 Norwegian language, where the "ae" ligature is the 27th letter of a 899 29-letter extended Latin alphabet. It is equivalent to the 28th 900 letter of the Swedish alphabet (also containing 29 letters), U+00E4 901 LATIN SMALL LETTER A WITH DIAERESIS, for which an "ae" cannot be 902 substituted according to current orthographic standards. That 903 character (U+00E4) is also part of the German alphabet where, unlike 904 in the Nordic languages, the two-character sequence "ae" is usually 905 treated as a fully acceptable alternate orthography for the "umlauted 906 a" character. The inverse is however not true, and those two 907 characters cannot necessarily be combined into an "umlauted a". This 908 also applies to another German character, the "umlauted o" (U+00F6 909 LATIN SMALL LETTER O WITH DIAERESIS) which, for example, cannot be 910 used for writing the name of the author "Goethe". It is also a 911 letter in the Swedish alphabet where, like the "a with diaeresis", it 912 cannot be correctly represented as "oe" and in the Norwegian 913 alphabet, where it is represented, not as "o with diaeresis", but as 914 "slashed o", U+00F8. 916 Some of the ligatures that have explicit code points in Unicode were 917 given special handling in IDNA2003 and now pose additional problems 918 in transition. See Section 7.2. 920 Additional cases with alphabets written right to left are described 921 in Section 4.5. 923 Matching and comparison algorithm selection often requires 924 information about the language being used, context, or both -- 925 information that is not available to IDNA or the DNS. Consequently, 926 these specifications make no attempt to treat combined characters in 927 any special way. A registry that is aware of the language context in 928 which labels are to be registered, and where that language sometimes 929 (or always) treats the two- character sequences as equivalent to the 930 combined form, should give serious consideration to applying a 931 "variant" model [RFC3743][RFC4290], or to prohibiting registration of 932 one of the forms entirely, to reduce the opportunities for user 933 confusion and fraud that would result from the related strings being 934 registered to different parties. 936 4.4. Case Mapping and Related Issues 938 In the DNS, ASCII letters are stored with their case preserved. 939 Matching during the query process is case-independent, but none of 940 the information that might be represented by choices of case has been 941 lost. That model has been accidentally helpful because, as people 942 have created DNS labels by catenating words (or parts of words) to 943 form labels, case has often been used to distinguish among components 944 and make the labels more memorable. 946 Since DNS servers do not get involved in parsing IDNs, they cannot do 947 case-independent matching. Thus, keeping the cases separate in 948 lookup or registration, and doing matching at the server, is not 949 feasible with IDNA or any similar approach. Case-matching must be 950 done, if desired, by IDN clients even though it wasn't done by ASCII- 951 only DNS clients. That situation was recognized in IDNA2003 and 952 nothing in these specifications fundamentally changes it or could do 953 so. In IDNA2003, all characters are case-folded and mapped by 954 clients in a standardized step. 956 Some characters do not have upper case forms. For example the 957 Unicode case folding operation maps Greek Final Form Sigma (U+03C2) 958 to the medial form (U+03C3) and maps Eszett (German Sharp S, U+00DF) 959 to "ss". Neither of these mappings is reversible because the upper 960 case of U+03C3 is the Upper Case Sigma (U+03A3) and "ss" is an ASCII 961 string. IDNA2008 permits, at the risk of some incompatibility, 962 slightly more flexibility in this area by avoiding case folding and 963 treating these characters as themselves. Approaches to handling one- 964 way mappings are discussed in Section 7.2. 966 Because IDNA2003 maps Final Sigma and Eszett to other characters, and 967 the reverse mapping is never possible, neither Final Sigma nor Eszett 968 can be represented in the ACE form of IDNA2003 IDN nor in the native 969 character (U-label) form derived from it. With IDNA2008, both 970 characters can be used in an IDN and so the A-label used for lookup 971 for any U-label containing those characters, is now different. See 972 Section 7.1 for a discussion of what kinds of changes might require 973 the IDNA prefix to change; after extended discussions, the WG came to 974 consensus that the change for these characters did not justify a 975 prefix change. 977 4.5. Right to Left Text 979 In order to be sure that the directionality of right to left text is 980 unambiguous, IDNA2003 required that any label in which right to left 981 characters appear both starts and ends with them and that it not 982 include any characters with strong left to right properties (that 983 excludes other alphabetic characters but permits European digits). 984 Any other string that contains a right to left character and does not 985 meet those requirements is rejected. This is one of the few places 986 where the IDNA algorithms (both in IDNA2003 and in IDAN2008) examine 987 an entire label, not just individual characters. The algorithmic 988 model used in IDNA2003 rejects the label when the final character in 989 a right to left string requires a combining mark in order to be 990 correctly represented. 992 That prohibition is not acceptable for writing systems for languages 993 written with consonantal alphabets to which diacritical vocalic 994 systems are applied, and for languages with orthographies derived 995 from them where the combining marks may have different functionality. 996 In both cases the combining marks can be essential components of the 997 orthography. Examples of this are Yiddish, written with an extended 998 Hebrew script, and Dhivehi (the official language of Maldives) which 999 is written in the Thaana script (which is, in turn, derived from the 1000 Arabic script). IDNA2008 removes the restriction on final combining 1001 characters with a new set of rules for right to left scripts and 1002 their characters. Those new rules are specified in [IDNA2008-Bidi]. 1004 5. IDNs and the Robustness Principle 1006 The "Robustness Principle" is often stated as "Be conservative about 1007 what you send and liberal in what you accept" (See, e.g., Section 1008 1.2.2 of the applications-layer Host Requirements specification 1009 [RFC1123]) This principle applies to IDNA. In applying the principle 1010 to registries as the source ("sender") of all registered and useful 1011 IDNs, registries are responsible for being conservative about what 1012 they register and put out in the Internet. For IDNs to work well, 1013 zone administrators (registries) must have and require sensible 1014 policies about what is registered -- conservative policies -- and 1015 implement and enforce them. 1017 Conversely, lookup applications are expected to reject labels that 1018 clearly violate global (protocol) rules (no one has ever seriously 1019 claimed that being liberal in what is accepted requires being 1020 stupid). However, once one gets past such global rules and deals 1021 with anything sensitive to script or locale, it is necessary to 1022 assume that garbage has not been placed into the DNS, i.e., one must 1023 be liberal about what one is willing to look up in the DNS rather 1024 than guessing about whether it should have been permitted to be 1025 registered. 1027 If a string cannot be successfully found in the DNS after the lookup 1028 processing described here, it makes no difference whether it simply 1029 wasn't registered or was prohibited by some rule at the registry. 1030 Application implementors should be aware that where DNS wildcards are 1031 used, the ability to successfully resolve a name does not guarantee 1032 that it was actually registered. 1034 6. Front-end and User Interface Processing for Lookup 1036 Domain names may be identified and processed in many contexts. They 1037 may be typed in by users either by themselves or embedded in an 1038 identifier such as email addresses, URIs, or IRIs. They may occur in 1039 running text or be processed by one system after being provided in 1040 another. Systems may try to normalize URLs to determine (or guess) 1041 whether a reference is valid or two references point to the same 1042 object without actually looking the objects up (comparison without 1043 lookup is necessary for URI types that are not intended to be 1044 resolved). Some of these goals may be more easily and reliably 1045 satisfied than others. While there are strong arguments for any 1046 domain name that is placed "on the wire" -- transmitted between 1047 systems -- to be in the zero-ambiguity forms of A-labels, it is 1048 inevitable that programs that process domain names will encounter 1049 U-labels or variant forms. 1051 An application that implements the IDNA protocol [IDNA2008-Protocol] 1052 will always take any user input and convert it to a set of Unicode 1053 code points. That user input may be acquired by any of several 1054 different input methods, all with differing conversion processes to 1055 be taken into consideration (e.g., typed on a keyboard, written by 1056 hand onto some sort of digitizer, spoken into a microphone and 1057 interpreted by a speech-to-text engine, etc.). The process of taking 1058 any particular user input and mapping it into a Unicode code point 1059 may be a simple one: If a user strikes the "A" key on a US English 1060 keyboard, without any modifiers such as the "Shift" key held down, in 1061 order to draw a Latin small letter A ("a"), many (perhaps most) 1062 modern operating system input methods will produce to the calling 1063 application the code point U+0061, encoded in a single octet. 1065 Sometimes the process is somewhat more complicated: a user might 1066 strike a particular set of keys to represent a combining macron 1067 followed by striking the "A" key in order to draw a Latin small 1068 letter A with a macron above it. Depending on the operating system, 1069 the input method chosen by the user, and even the parameters with 1070 which the application communicates with the input method, the result 1071 might be the code point U+0101 (encoded as two octets in UTF-8 or 1072 UTF-16, four octets in UTF-32, etc.), the code point U+0061 followed 1073 by the code point U+0304 (again, encoded in three or more octets, 1074 depending upon the encoding used) or even the code point U+FF41 1075 followed by the code point U+0304 (and encoded in some form). And 1076 these examples leave aside the issue of operating systems and input 1077 methods that do not use Unicode code points for their character set. 1079 In every case, applications (with the help of the operating systems 1080 on which they run and the input methods used) need to perform a 1081 mapping from user input into Unicode code points. 1083 The original version of the IDNA protocol [RFC3490] used a model 1084 whereby input was taken from the user, mapped (via whatever input 1085 method mechanisms were used) to a set of Unicode code points, and 1086 then further mapped to a set of Unicode code points using the 1087 Nameprep profile specified in [RFC3491]. In this procedure, there 1088 are two separate mapping steps: First, a mapping done by the input 1089 method (which might be controlled by the operating system, the 1090 application, or some combination) and then a second mapping performed 1091 by the Nameprep portion of the IDNA protocol. The mapping done in 1092 Nameprep includes a particular mapping table to re-map some 1093 characters to other characters, a particular normalization, and a set 1094 of prohibited characters. 1096 Note that the result of the two step mapping process means that the 1097 mapping chosen by the operating system or application in the first 1098 step might differ significantly from the mapping supplied by the 1099 Nameprep profile in the second step. This has advantages and 1100 disadvantages. Of course, the second mapping regularizes what gets 1101 looked up in the DNS, making for better interoperability between 1102 implementations which use the Nameprep mapping. However, the 1103 application or operating system may choose mappings in their input 1104 methods, which when passed through the second (Nameprep) mapping 1105 result in characters that are "surprising" to the end user. 1107 The other important feature of the original version of the IDNA 1108 protocol is that, with very few exceptions, it assumes that any set 1109 of Unicode code points provided to the Nameprep mapping can be mapped 1110 into a string of Unicode code points that are "sensible", even if 1111 that means mapping some code points to nothing (that is, removing the 1112 code points from the string). This allowed maximum flexibility in 1113 input strings. 1115 The present version of IDNA differs significantly in approach from 1116 the original version. First and foremost, it does not provide 1117 explicit mapping instructions. Instead, it assumes that the 1118 application (perhaps via an operating system input method) will do 1119 whatever mapping it requires to convert input into Unicode code 1120 points. This has the advantage of giving flexibility to the 1121 application to choose a mapping that is suitable for its user given 1122 specific user requirements, and avoids the two-step mapping of the 1123 original protocol. Instead of a mapping, the current version of IDNA 1124 provides a set of categories that can be used to specify the valid 1125 code points allowed in a domain name. 1127 In principle, an application ought to take user input of a domain 1128 name and convert it to the set of Unicode code points that represent 1129 the domain name the user intends. As a practical matter, of course, 1130 determining user intent is a tricky business, so an application needs 1131 to choose a reasonable mapping from user input. That may differ 1132 based on the particular circumstances of a user, depending on locale, 1133 language, type of input method, etc. It is up to the application to 1134 make a reasonable choice. 1136 7. Migration from IDNA2003 and Unicode Version Synchronization 1138 7.1. Design Criteria 1140 As mentioned above and in RFC 4690, two key goals of the IDNA2008 1141 design are 1143 o to enable applications to be agnostic about whether they are being 1144 run in environments supporting any Unicode version from 3.2 1145 onward, 1147 o to permit incrementally adding new characters, character groups, 1148 scripts, and other character collections as they are incorporated 1149 into Unicode, doing so without disruption and, in the long term, 1150 without "heavy" processes (an IETF consensus process is required 1151 by the IDNA2008 specifications and is expected to be required and 1152 used until significant experience accumulates with IDNA operations 1153 and new versions of Unicode). 1155 7.1.1. Summary and Discussion of IDNA Validity Criteria 1157 The general criteria for a label to be considered valid under IDNA 1158 are (the actual rules are rigorously defined in [IDNA2008-Protocol] 1159 and [IDNA2008-Tables]): 1161 o The characters are "letters", marks needed to form letters, 1162 numerals, or other code points used to write words in some 1163 language. Symbols, drawing characters, and various notational 1164 characters are intended to be permanently excluded. There is no 1165 evidence that they are important enough to Internet operations or 1166 internationalization to justify expansion of domain names beyond 1167 the general principle of "letters, digits, and hyphen". 1168 (Additional discussion and rationale for the symbol decision 1169 appears in Section 7.6). 1171 o Other than in very exceptional cases, e.g., where they are needed 1172 to write substantially any word of a given language, punctuation 1173 characters are excluded. The fact that a word exists is not proof 1174 that it should be usable in a DNS label and DNS labels are not 1175 expected to be usable for multiple-word phrases (although they are 1176 certainly not prohibited if the conventions and orthography of a 1177 particular language cause that to be possible). 1179 o Characters that are unassigned (have no character assignment at 1180 all) in the version of Unicode being used by the registry or 1181 application are not permitted, even on lookup. The issues 1182 involved in this decision are discussed in Section 7.7. 1184 o Any character that is mapped to another character by a current 1185 version of NFKC is prohibited as input to IDNA (for either 1186 registration or lookup). With a few exceptions, this principle 1187 excludes any character mapped to another by Nameprep [RFC3491]. 1189 The principles above drive the design of rules that are specified 1190 exactly in [IDNA2008-Tables]. Those rules identify the characters 1191 that are valid under IDNA. The rules themselves are normative, and 1192 the tables are derived from them, rather than vice versa. 1194 7.1.2. Labels in Registration 1196 Any label registered in a DNS zone must be validated -- i.e., the 1197 criteria for that label must be met -- in order for applications to 1198 work as intended. This principle is not new. For example, since the 1199 DNS was first deployed, zone administrators have been expected to 1200 verify that names meet "hostname" requirements [RFC0952] where those 1201 requirements are imposed by the expected applications. Other 1202 applications contexts, such as the later addition of special service 1203 location formats [RFC2782] imposed new requirements on zone 1204 administrators. For zones that will contain IDNs, support for 1205 Unicode version-independence requires restrictions on all strings 1206 placed in the zone. In particular, for such zones: 1208 o Any label that appears to be an A-label, i.e., any label that 1209 starts in "xn--", must be valid under IDNA, i.e., they must be 1210 valid A-labels, as discussed in Section 2 above. 1212 o The Unicode tables (i.e., tables of code points, character 1213 classes, and properties) and IDNA tables (i.e., tables of 1214 contextual rules such as those that appear in the Tables 1215 document), must be consistent on the systems performing or 1216 validating labels to be registered. Note that this does not 1217 require that tables reflect the latest version of Unicode, only 1218 that all tables used on a given system are consistent with each 1219 other. 1221 Under this model, registry tables will need to be updated (both the 1222 Unicode-associated tables and the tables of permitted IDN characters) 1223 to enable a new script or other set of new characters. The registry 1224 will not be affected by newer versions of Unicode, or newly- 1225 authorized characters, until and unless it wishes to support them. 1226 The zone administrator is responsible for verifying validity for IDNA 1227 as well as its local policies -- a more extensive set of checks than 1228 are required for looking up the labels. Systems looking up or 1229 resolving DNS labels, especially IDN DNS labels, must be able to 1230 assume that applicable registration rules were followed for names 1231 entered into the DNS. 1233 7.1.3. Labels in Lookup 1235 Anyone looking up a label in a DNS zone is required to 1237 o Maintain IDNA and Unicode tables that are consistent with regard 1238 to versions, i.e., unless the application actually executes the 1239 classification rules in [IDNA2008-Tables], its IDNA tables must be 1240 derived from the version of Unicode that is supported more 1241 generally on the system. As with registration, the tables need 1242 not reflect the latest version of Unicode but they must be 1243 consistent. 1245 o Validate the characters in labels to be looked up only to the 1246 extent of determining that the U-label does not contain 1247 "DISALLOWED" code points or code points that are unassigned in its 1248 version of Unicode. 1250 o Validate the label itself for conformance with a small number of 1251 whole-label rules. In particular, it must verify that 1253 * there are no leading combining marks, 1254 * the "bidi" conditions are met if right to left characters 1255 appear, 1257 * any required contextual rules are available, and 1259 * any contextual rules that are associated with Joiner Controls 1260 (and "CONTEXTJ" characters more generally) are tested. 1262 o Do not reject labels based on other contextual rules about 1263 characters, including mixed-script label prohibitions. Such rules 1264 may be used to influence presentation decisions in the user 1265 interface, but not to avoid looking up domain names. 1267 To further clarify the rules about handling characters that require 1268 contextual rules, note that one can have a context-required character 1269 (i.e., one that requires a rule), but no rule. In that case, the 1270 character is treated the same way DISALLOWED characters are treated, 1271 until and unless a rule is supplied. That state is more or less 1272 equivalent to "the idea of permitting this character is accepted in 1273 principle, but it won't be permitted in practice until consensus is 1274 reached on a safe way to use it". 1276 The ability to add a rule more or less exempts these characters from 1277 the prohibition against reclassifying characters from DISALLOWED to 1278 PVALID. 1280 And, obviously, "no rule" is different from "have a rule, but the 1281 test either succeeds or fails". 1283 Lookup applications that follow these rules, rather than having their 1284 own criteria for rejecting lookup attempts, are not sensitive to 1285 version incompatibilities with the particular zone registry 1286 associated with the domain name except for labels containing 1287 characters recently added to Unicode. 1289 An application or client that processes names according to this 1290 protocol and then resolves them in the DNS will be able to locate any 1291 name that is registered, as long as those registrations are valid 1292 under IDNA and its version of the IDNA tables is sufficiently up-to- 1293 date to interpret all of the characters in the label. Messages to 1294 users should distinguish between "label contains an unallocated code 1295 point" and other types of lookup failures. A failure on the basis of 1296 an old version of Unicode may lead the user to a desire to upgrade to 1297 a newer version, but will have no other ill effects (this is 1298 consistent with behavior in the transition to the DNS when some hosts 1299 could not yet handle some forms of names or record types). 1301 7.2. Changes in Character Interpretations 1303 As a consequence of the elimination of mapping, the current version 1304 of IDNA changes the interpretation of a few characters relative to 1305 its predecessors. This subsection outlines the issues and discusses 1306 possible transition strategies. 1308 7.2.1. Character Changes: Eszett and Final Sigma 1310 In those scripts that make case distinctions, there are a few 1311 characters for which an obvious and unique upper case character has 1312 not historically been available to match a lower case one or vice 1313 versa. For those characters, the mappings used in constructing the 1314 Stringprep tables for IDNA2003, performed using the Unicode CaseFold 1315 operation (See Section 5.8 of the Unicode Standard [Unicode51]), 1316 generate different characters or sets of characters. Those 1317 operations are not reversible and lose even more information than 1318 traditional upper case or lower case transformations, but are more 1319 useful than those transformations for comparison purposes. Two 1320 notable characters of this type are the German character Eszett 1321 (Sharp S, U+00DF) and the Greek Final Form Sigma (U+03C2). The 1322 former is case-folded to the ASCII string "ss", the latter to a 1323 medial (Lower Case) Sigma (U+03C3). 1325 7.2.2. Character Changes: Zero-Width Joiner and Non-Joiner 1327 IDNA2003 mapped both Zero-Width Joiner (ZWJ, U+200D) and Zero-Width 1328 Non-Joiner (ZWNJ, U+200C) to nothing, effectively dropping these 1329 characters from any label in which they appeared and treating strings 1330 containing them as identical to strings that did not. As discussed 1331 in Section 3.1.2 above, those characters are essential for writing 1332 many reasonable mnemonics for certain scripts. However, treating 1333 them as valid in the current version of IDNA, even with contextual 1334 restrictions, raises approximately the same problem as exists with 1335 Eszett and Final Sigma: strings that were valid under IDNA2003 have 1336 different interpretations as labels, and different A-labels, than the 1337 same strings under this newer version. 1339 7.2.3. Character Changes and the Need for Transition 1341 The decision to eliminate mandatory and standardized mappings, 1342 including case folding, from the IDNA2008 protocol in order to make 1343 A-labels and U-labels idempotent made these characters problematic. 1344 If they were to be disallowed, important words and mnemonics could 1345 not be written in orthographically reasonable ways. If they were to 1346 be permitted as distinct characters, there would be no information 1347 loss and registries would have more flexibility, but IDNA2003 and 1348 IDNA2008 lookups might result in different A-labels. 1350 With the understanding that there would be incompatibility either way 1351 but a judgment that the incompatibility was not significant enough to 1352 justify a prefix change, the WG concluded that Eszett and Final Form 1353 Sigma should be treated as distinct and Protocol-Valid characters. 1355 Since these characters are interpreted in different ways under the 1356 older and newer versions of IDNA, transition strategies and policies 1357 will be necessary. Some actions can reasonably be taken by 1358 applications client programs (those that perform lookup operations or 1359 cause them to be performed) but, because of the diversity of 1360 situations and uses of the DNS, much of the responsibility will need 1361 to fall on registries. 1363 Registries, especially those maintaining zones for third parties, 1364 must decide how to introduce a new service in a way that does not 1365 create confusion or significantly weaken or invalidate existing 1366 identifiers. This is not a new problem; registries were faced with 1367 similar issues when IDNs were introduced (potentially, and especially 1368 for Latin-based scripts, in conflict with existing labels that had 1369 been rendered in ASCII character by applying more or less 1370 standardized conventions) and when other new forms of strings have 1371 been permitted as labels. 1373 7.2.4. Transition Strategies 1375 There are several approaches to the introduction of new characters or 1376 changes in interpretation of existing characters from their mapped 1377 forms in the earlier version of IDNA. The transition issue is 1378 complicated because the forms of these labels after 1379 ToUnicode(ToASCII()) translation in IDNA2003 not only remain valid 1380 but do not provide strong indications of what the registrant 1381 intended: a string containing "ss" could have simply been intended to 1382 be that string or could have been intended to contain an Eszett, a 1383 string containing lower-case Sigma could have been intended to 1384 contain Final Sigma (one might make heuristic guesses based on 1385 position in a string, but the long tradition of forming labels by 1386 concatenating words makes such heuristics unreliable), and strings 1387 that do not contain ZWJ or ZWNJ might have been intended to contain 1388 them. Without any preference or claim to completeness, some of 1389 these, all of which have been used by registries in the past for 1390 similar transitions, are: 1392 1. Do not permit use of the newly-available character at the 1393 registry level. This might cause lookup failures if a domain 1394 name were to be written with the expectation of the IDNA2003 1395 mapping behavior, but would eliminate any possibility of false 1396 matches. 1398 2. Hold a "sunrise"-like arrangement in which holders of labels 1399 containing "ss" in the Eszett case, Lower Case Sigma in that 1400 case, or that might have contained ZWJ or ZWNJ in context, are 1401 given priority (and perhaps other benefits) for registering the 1402 corresponding string containing Eszett, Final Sigma, or the 1403 appropriate Zero-width character respectively. 1405 3. Adopt some sort of "variant" approach in which registrants obtain 1406 labels with both character forms. 1408 4. Adopt a different form of "variant" approach in which 1409 registration of additional strings that would produce the same 1410 A-label if interpreted according to IDNA2003 is either not 1411 permitted at all or permitted only by the registrant who already 1412 has one of the names. 1414 5. Ignore the issue and assume that the marketplace or other 1415 mechanisms will sort things out. 1417 In any event, a registry (at any level of the DNS tree) that chooses 1418 to permit labels to be registered that contains these characters, or 1419 considers doing so, will have to address the relationship with 1420 existing, possibly-conflicting, labels in some way, just as 1421 registries that already had a considerable number of labels did when 1422 IDNs were first introduced. 1424 7.3. Elimination of Character Mapping 1426 As discussed at length in Section 6, IDNA2003, via Nameprep (see 1427 Section 7.5), mapped many characters into related ones. Those 1428 mappings no longer exist as requirements in IDNA2008. These 1429 specifications strongly prefer that only A-labels or U-labels be used 1430 in protocol contexts and as much as practical more generally. 1431 IDNA2008 does anticipate situations in which some mapping at the time 1432 of user input into lookup applications is appropriate and desirable. 1433 The issues are discussed in Section 6 and specific recommendations 1434 are made in [IDNA2008-Mapping]. 1436 7.4. The Question of Prefix Changes 1438 The conditions that would have required a change in the IDNA ACE 1439 prefix ("xn--" for the version of IDNA specified in [RFC3490]) were 1440 of great concern to the community. A prefix change would have 1441 clearly been necessary if the algorithms were modified in a manner 1442 that would have created serious ambiguities during subsequent 1443 transition in registrations. This section summarizes the working 1444 group's conclusions about the conditions under which a change in the 1445 prefix would have been necessary and the implications of such a 1446 change. 1448 7.4.1. Conditions Requiring a Prefix Change 1450 An IDN prefix change would have been needed if a given string would 1451 be looked up or otherwise interpreted differently depending on the 1452 version of the protocol or tables being used. This IDNA upgrade 1453 would have required a prefix change if, and only if, one of the 1454 following four conditions were met: 1456 1. The conversion of an A-label to Unicode (i.e., a U-label) would 1457 have yielded one string under IDNA2003 (RFC3490) and a different 1458 string under IDNA2008. 1460 2. In a significant number of cases, an input string that was valid 1461 under IDNA2003 and also valid under IDNA2008 would have yielded 1462 two different A-labels with the different versions. This 1463 condition is believed to be essentially equivalent to the one 1464 above except for a very small number of edge cases that were not 1465 found to justify a prefix change (See Section 7.2). 1467 Note that if the input string was valid under one version and not 1468 valid under the other, this condition would not apply. See the 1469 first item in Section 7.4.2, below. 1471 3. A fundamental change was made to the semantics of the string that 1472 would be inserted in the DNS, e.g., if a decision were made to 1473 try to include language or script information in the encoding in 1474 addition to the string itself. 1476 4. A sufficiently large number of characters were added to Unicode 1477 so that the Punycode mechanism for block offsets would no longer 1478 reference the higher-numbered planes and blocks. This condition 1479 is unlikely even in the long term and certain not to arise in the 1480 next several years. 1482 7.4.2. Conditions Not Requiring a Prefix Change 1484 As a result of the principles described above, none of the following 1485 changes required a new prefix: 1487 1. Prohibition of some characters as input to IDNA. Such a 1488 prohibition might make names that were previously registered 1489 inaccessible, but did not change those names. 1491 2. Adjustments in IDNA tables or actions, including normalization 1492 definitions, that affected characters that were already invalid 1493 under IDNA2003. 1495 3. Changes in the style of the IDNA definition that did not alter 1496 the actions performed by IDNA. 1498 7.4.3. Implications of Prefix Changes 1500 While it might have been possible to make a prefix change, the costs 1501 of such a change are considerable. Registries could not have 1502 converted all IDNA2003 ("xn--") registrations to a new form at the 1503 same time and synchronize that change with applications supporting 1504 lookup. Unless all existing registrations were simply to be declared 1505 invalid (and perhaps even then) systems that needed to support both 1506 labels with old prefixes and labels with new ones would be required 1507 to first process a putative label under the IDNA2008 rules and try to 1508 look it up and then, if it were not found, would be required to 1509 process the label under IDNA2003 rules and look it up again. That 1510 process would probably have significantly slowed down all processing 1511 that involved IDNs in the DNS especially since a fully-qualified name 1512 might contain a mixture of labels that were registered with the old 1513 and new prefixes. That would have made DNS caching very difficult. 1514 In addition, looking up the same input string as two separate 1515 A-labels would have created some potential for confusion and attacks, 1516 since the labels could map to different targets and then resolve to 1517 different entries in the DNS. 1519 Consequently, a prefix change should have been, and was, avoided if 1520 at all possible, even if it means accepting some IDNA2003 decisions 1521 about character distinctions as irreversible and/or giving special 1522 treatment to edge cases. 1524 7.5. Stringprep Changes and Compatibility 1526 The Nameprep [RFC3491] specification, a key part of IDNA2003, is a 1527 profile of Stringprep [RFC3454]. While Nameprep is a Stringprep 1528 profile specific to IDNA, Stringprep is used by a number of other 1529 protocols. Were Stringprep to have been modified by IDNA2008, those 1530 changes to improve the handling of IDNs could cause problems for non- 1531 DNS uses, most notably if they affected identification and 1532 authentication protocols. Several elements of IDNA2008 give 1533 interpretations to strings prohibited under IDNA2003 or prohibit 1534 strings that IDNA2003 permitted. Those elements include the proposed 1535 new inclusion tables [IDNA2008-Tables], the reduction in the number 1536 of characters permitted as input for registration or lookup 1537 (Section 3), and even the proposed changes in handling of right to 1538 left strings [IDNA2008-Bidi]. IDNA2008 does not use Nameprep or 1539 Stringprep at all, so there are no side-effect changes to other 1540 protocols. 1542 It is particularly important to keep IDNA processing separate from 1543 processing for various security protocols because some of the 1544 constraints that are necessary for smooth and comprehensible use of 1545 IDNs may be unwanted or undesirable in other contexts. For example, 1546 the criteria for good passwords or passphrases are very different 1547 from those for desirable IDNs: passwords should be hard to guess, 1548 while domain names should normally be easily memorable. Similarly, 1549 internationalized SCSI identifiers and other protocol components are 1550 likely to have different requirements than IDNs. 1552 7.6. The Symbol Question 1554 One of the major differences between this specification and the 1555 original version of IDNA is that the original version permitted non- 1556 letter symbols of various sorts, including punctuation and line- 1557 drawing symbols, in the protocol. They were always discouraged in 1558 practice. In particular, both the "IESG Statement" about IDNA and 1559 all versions of the ICANN Guidelines specify that only language 1560 characters be used in labels. This specification disallows symbols 1561 entirely. There are several reasons for this, which include: 1563 1. As discussed elsewhere, the original IDNA specification assumed 1564 that as many Unicode characters as possible should be permitted, 1565 directly or via mapping to other characters, in IDNs. This 1566 specification operates on an inclusion model, extrapolating from 1567 the original "hostname" rules (LDH, see [IDNA2008-Defs]) -- which 1568 have served the Internet very well -- to a Unicode base rather 1569 than an ASCII base. 1571 2. Symbol names are more problematic than letters because there may 1572 be no general agreement on whether a particular glyph matches a 1573 symbol; there are no uniform conventions for naming; variations 1574 such as outline, solid, and shaded forms may or may not exist; 1575 and so on. As just one example, consider a "heart" symbol as it 1576 might appear in a logo that might be read as "I love...". While 1577 the user might read such a logo as "I love..." or "I heart...", 1578 considerable knowledge of the coding distinctions made in Unicode 1579 is needed to know that there is more than one "heart" character 1580 (e.g., U+2665, U+2661, and U+2765) and how to describe it. These 1581 issues are of particular importance if strings are expected to be 1582 understood or transcribed by the listener after being read out 1583 loud. 1585 3. Design of a screen reader used by blind Internet users who must 1586 listen to renderings of IDN domain names and possibly reproduce 1587 them on the keyboard becomes considerably more complicated when 1588 the names of characters are not obvious and intuitive to anyone 1589 familiar with the language in question. 1591 4. As a simplified example of this, assume one wanted to use a 1592 "heart" or "star" symbol in a label. This is problematic because 1593 those names are ambiguous in the Unicode system of naming (the 1594 actual Unicode names require far more qualification). A user or 1595 would-be registrant has no way to know -- absent careful study of 1596 the code tables -- whether it is ambiguous (e.g., where there are 1597 multiple "heart" characters) or not. Conversely, the user seeing 1598 the hypothetical label doesn't know whether to read it -- try to 1599 transmit it to a colleague by voice -- as "heart", as "love", as 1600 "black heart", or as any of the other examples below. 1602 5. The actual situation is even worse than this. There is no 1603 possible way for a normal, casual, user to tell the difference 1604 between the hearts of U+2665 and U+2765 and the stars of U+2606 1605 and U+2729 without somehow knowing to look for a distinction. We 1606 have a white heart (U+2661) and few black hearts. Consequently, 1607 describing a label as containing a heart is hopelessly ambiguous: 1608 we can only know that it contains one of several characters that 1609 look like hearts or have "heart" in their names. In cities where 1610 "Square" is a popular part of a location name, one might well 1611 want to use a square symbol in a label as well and there are far 1612 more squares of various flavors in Unicode than there are hearts 1613 or stars. 1615 The consequence of these ambiguities is that symbols are a very poor 1616 basis for reliable communication. Consistent with this conclusion, 1617 the Unicode standard recommends that strings used in identifiers not 1618 contain symbols or punctuation [Unicode-UAX31]. Of course, these 1619 difficulties with symbols do not arise with actual pictographic 1620 languages and scripts which would be treated like any other language 1621 characters; the two should not be confused. 1623 7.7. Migration Between Unicode Versions: Unassigned Code Points 1625 In IDNA2003, labels containing unassigned code points are looked up 1626 on the assumption that, if they appear in labels and can be mapped 1627 and then resolved, the relevant standards must have changed and the 1628 registry has properly allocated only assigned values. 1630 In the protocol described in these documents, strings containing 1631 unassigned code points must not be either looked up or registered. 1632 In summary, the status of an unassigned character with regard to the 1633 DISALLOWED, PROTOCOL-VALID, and CONTEXTUAL RULE REQUIRED categories 1634 cannot be evaluated until a character is actually assigned and known. 1635 There are several reasons for this, with the most important ones 1636 being: 1638 o Tests involving the context of characters (e.g., some characters 1639 being permitted only adjacent to others of specific types) and 1640 integrity tests on complete labels are needed. Unassigned code 1641 points cannot be permitted because one cannot determine whether 1642 particular code points will require contextual rules (and what 1643 those rules should be) before characters are assigned to them and 1644 the properties of those characters fully understood. 1646 o It cannot be known in advance, and with sufficient reliability, 1647 whether a newly-assigned code point will be associated with a 1648 character that would be disallowed by the rules in 1649 [IDNA2008-Tables] (such as a compatibility character). In 1650 IDNA2003, since there is no direct dependency on NFKC (many of the 1651 entries in Stringprep's tables are based on NFKC, but IDNA2003 1652 depends only on Stringprep), allocation of a compatibility 1653 character might produce some odd situations, but it would not be a 1654 problem. In IDNA2008, where compatibility characters are 1655 DISALLOWED unless character-specific exceptions are made, 1656 permitting strings containing unassigned characters to be looked 1657 up would violate the principle that characters in DISALLOWED are 1658 not looked up. 1660 o The Unicode Standard specifies that an unassigned code point 1661 normalizes (and, where relevant, case folds) to itself. If the 1662 code point is later assigned to a character, and particularly if 1663 the newly-assigned code point has a combining class that 1664 determines its placement relative to other combining characters, 1665 it could normalize to some other code point or sequence. 1667 It is possible to argue that the issues above are not important and 1668 that, as a consequence, it is better to retain the principle of 1669 looking up labels even if they contain unassigned characters because 1670 all of the important scripts and characters have been coded as of 1671 Unicode 5.1 and hence unassigned code points will be assigned only to 1672 obscure characters or archaic scripts. Unfortunately, that does not 1673 appear to be a safe assumption for at least two reasons. First, much 1674 the same claim of completeness has been made for earlier versions of 1675 Unicode. The reality is that a script that is obscure to much of the 1676 world may still be very important to those who use it. Cultural and 1677 linguistic preservation principles make it inappropriate to declare 1678 the script of no importance in IDNs. Second, we already have 1679 counterexamples in, e.g., the relationships associated with new Han 1680 characters being added (whether in the BMP or in Unicode Plane 2). 1682 Independent of the technical transition issues identified above, it 1683 can be observed that any addition of characters to an existing script 1684 to make it easier to use or to better accommodate particular 1685 languages may lead to transition issues. Such additions may change 1686 the preferred form for writing a particular string, changes that may 1687 be reflected, e.g., in keyboard transition modules that would 1688 necessarily be different from those for earlier versions of Unicode 1689 where the newer characters may not exist. This creates an inherent 1690 transition problem because attempts to access labels may use either 1691 the old or the new conventions, requiring registry action whether the 1692 older conventions were used in labels or not. The need to consider 1693 transition mechanisms is inherent to evolution of Unicode to better 1694 accommodate writing systems and is independent of how IDNs are 1695 represented in the DNS or how transitions among versions of those 1696 mechanisms occur. The requirement for transitions of this type is 1697 illustrated by the addition of Malayalam Chillu in Unicode 5.1.0. 1699 7.8. Other Compatibility Issues 1701 The 2003 IDNA model includes several odd artifacts of the context in 1702 which it was developed. Many, if not all, of these are potential 1703 avenues for exploits, especially if the registration process permits 1704 "source" names (names that have not been processed through IDNA and 1705 Nameprep) to be registered. As one example, since the character 1706 Eszett, used in German, is mapped by IDNA2003 into the sequence "ss" 1707 rather than being retained as itself or prohibited, a string 1708 containing that character but that is otherwise in ASCII is not 1709 really an IDN (in the U-label sense defined above) at all. After 1710 Nameprep maps the Eszett out, the result is an ASCII string and so 1711 does not get an xn-- prefix, but the string that can be displayed to 1712 a user appears to be an IDN. The newer version of the protocol 1713 eliminates this artifact. A character is either permitted as itself 1714 or it is prohibited; special cases that make sense only in a 1715 particular linguistic or cultural context can be dealt with as 1716 localization matters where appropriate. 1718 8. Name Server Considerations 1720 8.1. Processing Non-ASCII Strings 1722 Existing DNS servers do not know the IDNA rules for handling non- 1723 ASCII forms of IDNs, and therefore need to be shielded from them. 1724 All existing channels through which names can enter a DNS server 1725 database (for example, master files (as described in RFC 1034) and 1726 DNS update messages [RFC2136]) are IDN-unaware because they predate 1727 IDNA. Other sections of this document provide the needed shielding 1728 by ensuring that internationalized domain names entering DNS server 1729 databases through such channels have already been converted to their 1730 equivalent ASCII A-label forms. 1732 Because of the distinction made between the algorithms for 1733 Registration and Lookup in [IDNA2008-Protocol] (a domain name 1734 containing only ASCII codepoints cannot be converted to an A-label), 1735 there cannot be more than one A-label form for any given U-label. 1737 As specified in RFC 2181 [RFC2181], the DNS protocol explicitly 1738 allows domain labels to contain octets beyond the ASCII range 1739 (0000..007F), and this document does not change that. However, 1740 although the interpretation of octets 0080..00FF is well-defined in 1741 the DNS, many application protocols support only ASCII labels and 1742 there is no defined interpretation of these non-ASCII octets as 1743 characters and, in particular, no interpretation of case-independent 1744 matching for them (see, e.g., [RFC4343]). If labels containing these 1745 octets are returned to applications, unpredictable behavior could 1746 result. The A-label form, which cannot contain those characters, is 1747 the only standard representation for internationalized labels in the 1748 DNS protocol. 1750 8.2. Root and other DNS Server Considerations 1752 IDNs in A-label form will generally be somewhat longer than current 1753 domain names, so the bandwidth needed by the root servers is likely 1754 to go up by a small amount. Also, queries and responses for IDNs 1755 will probably be somewhat longer than typical queries historically, 1756 so EDNS0 [RFC2671] support may be more important (otherwise, queries 1757 and responses may be forced to go to TCP instead of UDP). 1759 9. Internationalization Considerations 1761 DNS labels and fully-qualified domain names provide mnemonics that 1762 assist in identifying and referring to resources on the Internet. 1763 IDNs expand the range of those mnemonics to include those based on 1764 languages and character sets other than Western European and Roman- 1765 derived ones. But domain "names" are not, in general, words in any 1766 language. The recommendations of the IETF policy on character sets 1767 and languages, (BCP 18 [RFC2277]) are applicable to situations in 1768 which language identification is used to provide language-specific 1769 contexts. The DNS is, by contrast, global and international and 1770 ultimately has nothing to do with languages. Adding languages (or 1771 similar context) to IDNs generally, or to DNS matching in particular, 1772 would imply context dependent matching in DNS, which would be a very 1773 significant change to the DNS protocol itself. It would also imply 1774 that users would need to identify the language associated with a 1775 particular label in order to look that label up. That knowledge is 1776 generally not available because many labels are not words in any 1777 language and some may be words in more than one. 1779 10. IANA Considerations 1781 This section gives an overview of IANA registries required for IDNA. 1782 The actual definitions of, and specifications for, the first two, 1783 which must be newly-created for IDNA2008, appear in 1784 [IDNA2008-Tables]. This document describes the registries but does 1785 not specify any IANA actions. 1787 10.1. IDNA Character Registry 1789 The distinction among the major categories "UNASSIGNED", 1790 "DISALLOWED", "PROTOCOL-VALID", and "CONTEXTUAL RULE REQUIRED" is 1791 made by special categories and rules that are integral elements of 1792 [IDNA2008-Tables]. While not normative, an IANA registry of 1793 characters and scripts and their categories, updated for each new 1794 version of Unicode and the characters it contains, will be convenient 1795 for programming and validation purposes. The details of this 1796 registry are specified in [IDNA2008-Tables]. 1798 10.2. IDNA Context Registry 1800 IANA will create and maintain a list of approved contextual rules for 1801 characters that are defined in the IDNA Character Registry list as 1802 requiring a Contextual Rule (i.e., the types of rule described in 1803 Section 3.1.2). The details for those rules appear in 1804 [IDNA2008-Tables]. 1806 10.3. IANA Repository of IDN Practices of TLDs 1808 This registry, historically described as the "IANA Language Character 1809 Set Registry" or "IANA Script Registry" (both somewhat misleading 1810 terms) is maintained by IANA at the request of ICANN. It is used to 1811 provide a central documentation repository of the IDN policies used 1812 by top level domain (TLD) registries who volunteer to contribute to 1813 it and is used in conjunction with ICANN Guidelines for IDN use. 1815 It is not an IETF-managed registry and, while the protocol changes 1816 specified here may call for some revisions to the tables, these 1817 specifications have no direct effect on that registry and no IANA 1818 action is required as a result. 1820 11. Security Considerations 1822 11.1. General Security Issues with IDNA 1824 This document is purely explanatory and informational and 1825 consequently introduces no new security issues. It would, of course, 1826 be a poor idea for someone to try to implement from it; such an 1827 attempt would almost certainly lead to interoperability problems and 1828 might lead to security ones. A discussion of security issues with 1829 IDNA, including some relevant history, appears in [IDNA2008-Defs]. 1831 12. Acknowledgments 1833 The editor and contributors would like to express their thanks to 1834 those who contributed significant early (pre-WG) review comments, 1835 sometimes accompanied by text, Paul Hoffman, Simon Josefsson, and Sam 1836 Weiler. In addition, some specific ideas were incorporated from 1837 suggestions, text, or comments about sections that were unclear 1838 supplied by Vint Cerf, Frank Ellerman, Michael Everson, Asmus 1839 Freytag, Erik van der Poel, Michel Suignard, and Ken Whistler. 1840 Thanks are also due to Vint Cerf, Lisa Dusseault, Debbie Garside, and 1841 Jefsey Morfin for conversations that led to considerable improvements 1842 in the content of this document and to several others, including Ben 1843 Campbell, Martin Duerst, Subramanian Moonesamy, Peter Saint-Andre, 1844 and Dan Winship, for catching specific errors and recommending 1845 corrections. 1847 A meeting was held on 30 January 2008 to attempt to reconcile 1848 differences in perspective and terminology about this set of 1849 specifications between the design team and members of the Unicode 1850 Technical Consortium. The discussions at and subsequent to that 1851 meeting were very helpful in focusing the issues and in refining the 1852 specifications. The active participants at that meeting were (in 1853 alphabetic order as usual) Harald Alvestrand, Vint Cerf, Tina Dam, 1854 Mark Davis, Lisa Dusseault, Patrik Faltstrom (by telephone), Cary 1855 Karp, John Klensin, Warren Kumari, Lisa Moore, Erik van der Poel, 1856 Michel Suignard, and Ken Whistler. We express our thanks to Google 1857 for support of that meeting and to the participants for their 1858 contributions. 1860 Useful comments and text on the WG versions of the draft were 1861 received from many participants in the IETF "IDNABIS" WG and a number 1862 of document changes resulted from mailing list discussions made by 1863 that group. Marcos Sanz provided specific analysis and suggestions 1864 that were exceptionally helpful in refining the text, as did Vint 1865 Cerf, Martin Duerst, Andrew Sullivan, and Ken Whistler. Lisa 1866 Dusseault provided extensive editorial suggestions during the spring 1867 of 2009, most of which were incorporated. 1869 13. Contributors 1871 While the listed editor held the pen, the core of this document and 1872 the initial WG version represents the joint work and conclusions of 1873 an ad hoc design team consisting of the editor and, in alphabetic 1874 order, Harald Alvestrand, Tina Dam, Patrik Faltstrom, and Cary Karp. 1875 Considerable material describing mapping principles has been 1876 incorporated from a draft of [IDNA2008-Mapping] by Pete Resnick and 1877 Paul Hoffman. In addition, there were many specific contributions 1878 and helpful comments from those listed in the Acknowledgments section 1879 and others who have contributed to the development and use of the 1880 IDNA protocols. 1882 14. References 1884 14.1. Normative References 1886 [ASCII] American National Standards Institute (formerly United 1887 States of America Standards Institute), "USA Code for 1888 Information Interchange", ANSI X3.4-1968, 1968. 1890 ANSI X3.4-1968 has been replaced by newer versions with 1891 slight modifications, but the 1968 version remains 1892 definitive for the Internet. 1894 [IDNA2008-Bidi] 1895 Alvestrand, H. and C. Karp, "An updated IDNA criterion for 1896 right to left scripts", August 2009, . 1899 [IDNA2008-Defs] 1900 Klensin, J., "Internationalized Domain Names for 1901 Applications (IDNA): Definitions and Document Framework", 1902 August 2009, . 1905 [IDNA2008-Protocol] 1906 Klensin, J., "Internationalized Domain Names in 1907 Applications (IDNA): Protocol", August 2009, . 1910 [IDNA2008-Tables] 1911 Faltstrom, P., "The Unicode Code Points and IDNA", 1912 August 2009, . 1915 A version of this document is available in HTML format at 1916 http://stupid.domain.name/idnabis/ 1917 draft-ietf-idnabis-tables-06.html 1919 [RFC3490] Faltstrom, P., Hoffman, P., and A. Costello, 1920 "Internationalizing Domain Names in Applications (IDNA)", 1921 RFC 3490, March 2003. 1923 [RFC3492] Costello, A., "Punycode: A Bootstring encoding of Unicode 1924 for Internationalized Domain Names in Applications 1925 (IDNA)", RFC 3492, March 2003. 1927 [Unicode-UAX15] 1928 The Unicode Consortium, "Unicode Standard Annex #15: 1929 Unicode Normalization Forms", March 2008, 1930 . 1932 [Unicode51] 1933 The Unicode Consortium, "The Unicode Standard, Version 1934 5.1.0", 2008. 1936 defined by: The Unicode Standard, Version 5.0, Boston, MA, 1937 Addison-Wesley, 2007, ISBN 0-321-48091-0, as amended by 1938 Unicode 5.1.0 1939 (http://www.unicode.org/versions/Unicode5.1.0/). 1941 14.2. Informative References 1943 [BIG5] Institute for Information Industry of Taiwan, "Computer 1944 Chinese Glyph and Character Code Mapping Table, Technical 1945 Report C-26", 1984. 1947 There are several forms and variations and a closely- 1948 related standard, CNS 11643. See the discussion in 1949 Chapter 3 of Lunde, K., CJKV Information Processing, 1950 O'Reilly & Associates, 1999 1952 [GB18030] "Chinese National Standard GB 18030-2000: Information 1953 Technology -- Chinese ideograms coded character set for 1954 information interchange -- Extension for the basic set.", 1955 2000. 1957 [IDNA2008-Mapping] 1958 Resnick, P., "Mapping Characters in IDNA", August 2009, . 1962 [RFC0810] Feinler, E., Harrenstien, K., Su, Z., and V. White, "DoD 1963 Internet host table specification", RFC 810, March 1982. 1965 [RFC0952] Harrenstien, K., Stahl, M., and E. Feinler, "DoD Internet 1966 host table specification", RFC 952, October 1985. 1968 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 1969 STD 13, RFC 1034, November 1987. 1971 [RFC1035] Mockapetris, P., "Domain names - implementation and 1972 specification", STD 13, RFC 1035, November 1987. 1974 [RFC1123] Braden, R., "Requirements for Internet Hosts - Application 1975 and Support", STD 3, RFC 1123, October 1989. 1977 [RFC2136] Vixie, P., Thomson, S., Rekhter, Y., and J. Bound, 1978 "Dynamic Updates in the Domain Name System (DNS UPDATE)", 1979 RFC 2136, April 1997. 1981 [RFC2181] Elz, R. and R. Bush, "Clarifications to the DNS 1982 Specification", RFC 2181, July 1997. 1984 [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and 1985 Languages", BCP 18, RFC 2277, January 1998. 1987 [RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", 1988 RFC 2671, August 1999. 1990 [RFC2673] Crawford, M., "Binary Labels in the Domain Name System", 1991 RFC 2673, August 1999. 1993 [RFC2782] Gulbrandsen, A., Vixie, P., and L. Esibov, "A DNS RR for 1994 specifying the location of services (DNS SRV)", RFC 2782, 1995 February 2000. 1997 [RFC3454] Hoffman, P. and M. Blanchet, "Preparation of 1998 Internationalized Strings ("stringprep")", RFC 3454, 1999 December 2002. 2001 [RFC3491] Hoffman, P. and M. Blanchet, "Nameprep: A Stringprep 2002 Profile for Internationalized Domain Names (IDN)", 2003 RFC 3491, March 2003. 2005 [RFC3743] Konishi, K., Huang, K., Qian, H., and Y. Ko, "Joint 2006 Engineering Team (JET) Guidelines for Internationalized 2007 Domain Names (IDN) Registration and Administration for 2008 Chinese, Japanese, and Korean", RFC 3743, April 2004. 2010 [RFC3987] Duerst, M. and M. Suignard, "Internationalized Resource 2011 Identifiers (IRIs)", RFC 3987, January 2005. 2013 [RFC4290] Klensin, J., "Suggested Practices for Registration of 2014 Internationalized Domain Names (IDN)", RFC 4290, 2015 December 2005. 2017 [RFC4343] Eastlake, D., "Domain Name System (DNS) Case Insensitivity 2018 Clarification", RFC 4343, January 2006. 2020 [RFC4690] Klensin, J., Faltstrom, P., Karp, C., and IAB, "Review and 2021 Recommendations for Internationalized Domain Names 2022 (IDNs)", RFC 4690, September 2006. 2024 [RFC4713] Lee, X., Mao, W., Chen, E., Hsu, N., and J. Klensin, 2025 "Registration and Administration Recommendations for 2026 Chinese Domain Names", RFC 4713, October 2006. 2028 [Unicode-Security] 2029 The Unicode Consortium, "Unicode Technical Standard #39: 2030 Unicode Security Mechanisms", August 2008, 2031 . 2033 [Unicode-UAX31] 2034 The Unicode Consortium, "Unicode Standard Annex #31: 2035 Unicode Identifier and Pattern Syntax", March 2008, 2036 . 2038 [Unicode-UTR36] 2039 The Unicode Consortium, "Unicode Technical Report #36: 2040 Unicode Security Considerations", July 2008, 2041 . 2043 Appendix A. Change Log 2045 [[ RFC Editor: Please remove this appendix. ]] 2047 A.1. Changes between Version -00 and Version -01 of 2048 draft-ietf-idnabis-rationale 2050 o Clarified the U-label definition to note that U-labels must 2051 contain at least one non-ASCII character. Also clarified the 2052 relationship among label types. 2054 o Rewrote the discussion of Labels in Registration (Section 7.1.2) 2055 and related text about IDNA-validity (in the "Defs" document as of 2056 -04 of this one) to narrow its focus and remove more general 2057 restrictions. Added a temporary note in line to explain the 2058 situation. 2060 o Changed the "IDNA uses Unicode" statement to focus on 2061 compatibility with IDNA2003 and avoid more general or 2062 controversial assertions. 2064 o Added a discussion of examples to Section 7.1 2066 o Made a number of other small editorial changes and corrections 2067 suggested by Mark Davis. 2069 o Added several more discussion anchors and notes and expanded or 2070 updated some existing ones. 2072 A.2. Version -02 2074 o Trimmed change log, removing information about pre-WG drafts. 2076 o Adjusted discussion of Contextual Rules to match the new location 2077 of the tables and some conceptual material. 2079 o Rewrote the material on preprocessing somewhat. 2081 o Moved the material about relationships with IDNA2003 to be part of 2082 a single section on transitions. 2084 o Removed several placeholders and made editorial changes in 2085 accordance with decisions made at IETF 72 in Dublin and not 2086 disputed on the mailing list. 2088 A.3. Version -03 2090 This special update to the Rationale document is intended to try to 2091 get the discussion of what is normative or not under control. While 2092 the IETF does not normally annotate individual sections of documents 2093 with whether they are normative or not, concerns that we don't know 2094 which is which, claims that some material is normative that would be 2095 problematic if so classified, etc., argue that we should at least be 2096 able to have a clear discussion on the subject. 2098 Two annotations have been applied to sections that might reasonably 2099 be considered normative. One annotation is based on the list of 2100 sections in Mark Davis's note of 29 September (http:// 2101 www.alvestrand.no/pipermail/idna-update/2008-September/002667.html). 2102 The other is based on an elaboration of John Klensin's response on 7 2103 October (http://www.alvestrand.no/pipermail/idna-update/2008-October/ 2104 002691.html). These should just be considered two suggestions to 2105 illuminate and, one hopes, advance the Working Group's discussions. 2107 Some additional editorial changes have been made, but they are 2108 basically trivial. In the editor's judgment, it is not possible to 2109 make significantly more progress with this document until the matter 2110 of document organization is settled. 2112 A.4. Version -04 2114 o Definitional and other normative material moved to new document 2115 (draft-ietf-idnabis-defs). Version -03 annotations removed. 2117 o Material on differences between IDNA2003 and IDNA2008 moved to an 2118 appendix in Protocol. 2120 o Material left over from the origins of this document as a 2121 preliminary proposal has been removed or rewritten. 2123 o Changes made to reflect consensus call results, including removing 2124 several placeholder notes for discussion. 2126 o Added more material, including discussion of historic scripts, to 2127 Section 3.2 on registration policies. 2129 o Added a new section (Section 7.2) to contain specific discussion 2130 of handling of characters that are interpreted differently in 2131 input to IDNA2003 and 2008. 2133 o Some material, including this section/appendix, rearranged. 2135 A.5. Version -05 2137 o Many small editorial changes, including changes to eliminate the 2138 last vestiges of what appeared to be 2119 language (upper-case 2139 MUST, SHOULD, or MAY) and small adjustments to terminology. 2141 A.6. Version -06 2143 o Removed Security Considerations material and pointed to Defs, 2144 where it now appears as of version 05. 2146 o Started changing uses of "IDNA2008" in running text to "in these 2147 specifications" or the equivalent. These documents are titled 2148 simply "IDNA"; once they are standardized, "the current version" 2149 may be a more appropriate reference than one containing a year. 2150 As discussed on the mailing list, we can and should discuss how to 2151 refer to these documents at an appropriate time (e.g., when we 2152 know when we will be finished) but, in the interim, it seems 2153 appropriate to simply start getting rid of the version-specific 2154 terminology where it can naturally be removed. 2156 o Additional discussion of mappings, etc., especially for case- 2157 sensitivity. 2159 o Clarified relationship to base DNS specifications. 2161 o Consolidated discussion of lookup of unassigned characters. 2163 o More editorial fine-tuning. 2165 A.7. Version -07 2167 o Revised terminology by adding terms: NR-LDH-label, Invalid-A-label 2168 (or False-A-label), R-LDH-label, valid IDNA-label in 2169 Section 1.3.2. 2171 o Moved the "name server considerations" material to this document 2172 from Protocol because it is non-normative and not part of the 2173 protocol itself. 2175 o To improve clarity, redid discussion of the reasons why looking up 2176 unassigned code points is prohibited. 2178 o Editorial and other non-substantive corrections to reflect earlier 2179 errors as well as new definitions and terminology. 2181 A.8. Version -08 2183 o Slight revision to "contextual" discussion (Section 3.1.2) and 2184 moving it to a separate subsection, rather than under "PVALID", 2185 for better parallelism with Tables. Also reflected Mark's 2186 comments about the limitations of the approach. 2188 o Added placeholder notes as reminders of where references to the 2189 other documents need Section numbers. More of these will be added 2190 as needed (feel free to identify relevant places), but the actual 2191 section numbers will not be inserted until the documents are 2192 completely stable, i.e., on their way to the RFC Editor. 2194 A.9. Version -09 2196 o Small editorial changes to clarify transition possibilities. 2198 o Small clarification to the description of DNS "exact match". 2200 o Added discussion of adding characters to an existing script to the 2201 discussion of unassigned code point transitions in Section 7.7. 2203 o Tightened up the discussion of non-ASCII string processing 2204 (Section 8.1) slightly. 2206 o Removed some placeholders and comments that have been around long 2207 enough to be considered acceptable or that no longer seem 2208 necessary for other reasons. 2210 A.10. Version -10 2212 o Extensive editorial improvements, mostly due to suggestions from 2213 Lisa Dusseault. 2215 o Changes required for the new "mapping" approach and document have, 2216 in general, not been incorporated despite several suggestions. 2217 The editor intends to wait until the mapping model is stable, or 2218 at least until -11 of this document, before trying to incorporate 2219 those suggestions. 2221 A.11. Version -11 2223 o Several placeholders for additional material or editing have been 2224 removed since no comments have been received. 2226 o Updated references. 2228 o Corrected an apparent patching error in Section 1.6 and another 2229 one in Section 4.3. 2231 o Adjusted several sections that had not properly reflected removal 2232 of the material that is now in the Definitions document and 2233 removed an unnecessary one. 2235 o New material added to Section 3.2 about registration policy issues 2236 to reflect discussions on the mailing list. 2238 o Incorporated mapping material from the former "Architectural 2239 Principles" of version -01 of the Mapping draft into Section 6 and 2240 removed most of the prior mapping material and explanations. 2242 o Eliminated the former Section 7.3 ("More Flexibility in User 2243 Agents"), moving its material into Section 4.2. The replacement 2244 section is basically a placeholder to retain the mapping issues as 2245 one of the migration topics. Note that this item and the previous 2246 one involve considerable text, so people should check things 2247 carefully. 2249 o Corrected several typographical and editorial errors that don't 2250 fall into any of the above categories. 2252 A.12. Version -12 2254 o Got rid of the term "IDNA-valid". It no longer appears in 2255 Definitions and we didn't really need the extra term. Where the 2256 concept was needed, the text now says "valid under IDNA" or 2257 equivalent. 2259 o Adjusted Acknowledgments to remove Mark Davis's name, per his 2260 request and advice from IETF Trust Counsel. 2262 o Incorporated other changes from WG Last Call. 2264 o Small typographical and editorial corrections. 2266 A.13. Version -13 2268 o Substituted in Section numbers to references to other IDNA2008 2269 documents. 2271 A.14. Version -14 2273 This is the version of the document produced to reflect comments on 2274 IETF Last Call. For the convenience of those who made comments and 2275 of the IESG in evaluating them, this section therefore identifies 2276 non-editorial changes made in response to Last Call comments in 2277 somewhat more detail than may be usual. 2279 o Removed the discussion of DNSSEC after extensive discussion on the 2280 IETF and IDNABIS lists. 2282 o Modified the discussion of prefix changes to make it clear that 2283 the decisions have been made, rather than still representing open 2284 issues. (Dan Winship review, 20091013) 2286 o Suggested explicit identification of domain name slots in 2287 protocols that use IDNA. Peter Saint-Andre, 20091019. 2289 o Several other clarifications as suggested by Peter Saint-Andre, 2290 20091019. 2292 o Several minor editorial corrections per suggestions in Ben 2293 Campbell's Gen-ART review 20091013. 2295 o Typo corrections. 2297 A.15. Version -15 2299 o Rewrote and expanded the "transition" material of Section 7.2. 2301 Author's Address 2303 John C Klensin 2304 1770 Massachusetts Ave, Ste 322 2305 Cambridge, MA 02140 2306 USA 2308 Phone: +1 617 245 1457 2309 Email: john+ietf@jck.com