idnits 2.17.00 (12 Aug 2021) /tmp/idnits8423/draft-ietf-idnabis-rationale-10.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** The document seems to lack a License Notice according IETF Trust Provisions of 28 Dec 2009, Section 6.b.i or Provisions of 12 Sep 2009 Section 6.b -- however, there's a paragraph with a matching beginning. Boilerplate error? (You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Feb 2009 rather than one of the newer Notices. See https://trustee.ietf.org/license-info/.) Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 81: '... CONTEXTUAL RULE REQUIRED . . . . . . ...' RFC 2119 keyword, line 460: '... CONTEXTUAL RULE REQUIRED, DISALLOWED ...' RFC 2119 keyword, line 486: '...3.1.2. CONTEXTUAL RULE REQUIRED...' RFC 2119 keyword, line 1572: '...nd CONTEXTUAL RULE REQUIRED categories...' RFC 2119 keyword, line 1700: '...esented directly in ASCII MUST use the...' (2 more instances...) 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 (June 18, 2009) is 4719 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ASCII' is defined on line 1847, but no explicit reference was found in the text == Unused Reference: 'Unicode-UAX15' is defined on line 1888, but no explicit reference was found in the text == Unused Reference: 'BIG5' is defined on line 1904, but no explicit reference was found in the text == Unused Reference: 'GB18030' is defined on line 1913, but no explicit reference was found in the text == Unused Reference: 'RFC0810' is defined on line 1918, but no explicit reference was found in the text == Unused Reference: 'RFC2673' is defined on line 1949, but no explicit reference was found in the text == Unused Reference: 'RFC4690' is defined on line 1979, but no explicit reference was found in the text == Unused Reference: 'Unicode-UTR36' is defined on line 1997, but no explicit reference was found in the text == Outdated reference: draft-ietf-idnabis-defs has been published as RFC 5890 == Outdated reference: draft-ietf-idnabis-tables has been published as RFC 5892 ** Obsolete normative reference: RFC 3490 (Obsoleted by RFC 5890, RFC 5891) -- Obsolete informational reference (is this intentional?): RFC 810 (Obsoleted by RFC 952) -- Obsolete informational reference (is this intentional?): RFC 2535 (Obsoleted by RFC 4033, RFC 4034, RFC 4035) -- Obsolete informational reference (is this intentional?): RFC 2671 (Obsoleted by RFC 6891) -- Obsolete informational reference (is this intentional?): RFC 2673 (Obsoleted by RFC 6891) -- Obsolete informational reference (is this intentional?): RFC 3454 (Obsoleted by RFC 7564) -- Obsolete informational reference (is this intentional?): RFC 3491 (Obsoleted by RFC 5891) Summary: 3 errors (**), 0 flaws (~~), 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 June 18, 2009 4 Intended status: Informational 5 Expires: December 20, 2009 7 Internationalized Domain Names for Applications (IDNA): Background, 8 Explanation, and Rationale 9 draft-ietf-idnabis-rationale-10.txt 11 Status of this Memo 13 This Internet-Draft is submitted to IETF in full conformance with the 14 provisions of BCP 78 and BCP 79. This document may contain material 15 from IETF Documents or IETF Contributions published or made publicly 16 available before November 10, 2008. The person(s) controlling the 17 copyright in some of this material may not have granted the IETF 18 Trust the right to allow modifications of such material outside the 19 IETF Standards Process. Without obtaining an adequate license from 20 the person(s) controlling the copyright in such materials, this 21 document may not be modified outside the IETF Standards Process, and 22 derivative works of it may not be created outside the IETF Standards 23 Process, except to format it for publication as an RFC or to 24 translate it into languages other than English. 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 December 20, 2009. 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 in effect on the date of 51 publication of this document (http://trustee.ietf.org/license-info). 52 Please review these documents carefully, as they describe your rights 53 and restrictions with respect to this document. 55 Abstract 57 Several years have passed since the original protocol for 58 Internationalized Domain Names (IDNs) was completed and deployed. 59 During that time, a number of issues have arisen, including the need 60 to update the system to deal with newer versions of Unicode. Some of 61 these issues require tuning of the existing protocols and the tables 62 on which they depend. This document provides an overview of a 63 revised system and provides explanatory material for its components. 65 Table of Contents 67 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 68 1.1. Context and Overview . . . . . . . . . . . . . . . . . . . 4 69 1.2. Discussion Forum . . . . . . . . . . . . . . . . . . . . . 5 70 1.3. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 71 1.3.1. Documents and Standards . . . . . . . . . . . . . . . 5 72 1.3.2. DNS "Name" Terminology . . . . . . . . . . . . . . . . 5 73 1.3.3. New Terminology and Restrictions . . . . . . . . . . . 6 74 1.4. Objectives . . . . . . . . . . . . . . . . . . . . . . . . 7 75 1.5. Applicability and Function of IDNA . . . . . . . . . . . . 7 76 1.6. Comprehensibility of IDNA Mechanisms and Processing . . . 8 77 2. Processing in IDNA2008 . . . . . . . . . . . . . . . . . . . . 9 78 3. Permitted Characters: An Inclusion List . . . . . . . . . . . 9 79 3.1. A Tiered Model of Permitted Characters and Labels . . . . 10 80 3.1.1. PROTOCOL-VALID . . . . . . . . . . . . . . . . . . . . 10 81 3.1.2. CONTEXTUAL RULE REQUIRED . . . . . . . . . . . . . . . 11 82 3.1.2.2. Rules and Their Application . . . . . . . . . . . 12 83 3.1.3. DISALLOWED . . . . . . . . . . . . . . . . . . . . . . 12 84 3.1.4. UNASSIGNED . . . . . . . . . . . . . . . . . . . . . . 13 85 3.2. Registration Policy . . . . . . . . . . . . . . . . . . . 13 86 3.3. Layered Restrictions: Tables, Context, Registration, 87 Applications . . . . . . . . . . . . . . . . . . . . . . . 14 88 4. Issues that Constrain Possible Solutions . . . . . . . . . . . 15 89 4.1. Display and Network Order . . . . . . . . . . . . . . . . 15 90 4.2. Entry and Display in Applications . . . . . . . . . . . . 16 91 4.3. Linguistic Expectations: Ligatures, Digraphs, and 92 Alternate Character Forms . . . . . . . . . . . . . . . . 17 93 4.4. Case Mapping and Related Issues . . . . . . . . . . . . . 18 94 4.5. Right to Left Text . . . . . . . . . . . . . . . . . . . . 19 95 5. IDNs and the Robustness Principle . . . . . . . . . . . . . . 20 96 6. Front-end and User Interface Processing for Lookup . . . . . . 20 97 7. Migration from IDNA2003 and Unicode Version Synchronization . 24 98 7.1. Design Criteria . . . . . . . . . . . . . . . . . . . . . 24 99 7.1.1. Summary and Discussion of IDNA Validity Criteria . . . 24 100 7.1.2. Labels in Registration . . . . . . . . . . . . . . . . 25 101 7.1.3. Labels in Lookup . . . . . . . . . . . . . . . . . . . 26 102 7.2. Changes in Character Interpretations . . . . . . . . . . . 27 103 7.3. More Flexibility in User Agents . . . . . . . . . . . . . 28 104 7.4. The Question of Prefix Changes . . . . . . . . . . . . . . 30 105 7.4.1. Conditions Requiring a Prefix Change . . . . . . . . . 30 106 7.4.2. Conditions Not Requiring a Prefix Change . . . . . . . 31 107 7.4.3. Implications of Prefix Changes . . . . . . . . . . . . 31 108 7.5. Stringprep Changes and Compatibility . . . . . . . . . . . 31 109 7.6. The Symbol Question . . . . . . . . . . . . . . . . . . . 32 110 7.7. Migration Between Unicode Versions: Unassigned Code 111 Points . . . . . . . . . . . . . . . . . . . . . . . . . . 33 112 7.8. Other Compatibility Issues . . . . . . . . . . . . . . . . 35 113 8. Name Server Considerations . . . . . . . . . . . . . . . . . . 35 114 8.1. Processing Non-ASCII Strings . . . . . . . . . . . . . . . 36 115 8.2. DNSSEC Authentication of IDN Domain Names . . . . . . . . 36 116 8.3. Root and other DNS Server Considerations . . . . . . . . . 37 117 9. Internationalization Considerations . . . . . . . . . . . . . 37 118 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 37 119 10.1. IDNA Character Registry . . . . . . . . . . . . . . . . . 38 120 10.2. IDNA Context Registry . . . . . . . . . . . . . . . . . . 38 121 10.3. IANA Repository of IDN Practices of TLDs . . . . . . . . . 38 122 11. Security Considerations . . . . . . . . . . . . . . . . . . . 38 123 11.1. General Security Issues with IDNA . . . . . . . . . . . . 38 124 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 39 125 13. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 39 126 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 40 127 14.1. Normative References . . . . . . . . . . . . . . . . . . . 40 128 14.2. Informative References . . . . . . . . . . . . . . . . . . 41 129 Appendix A. Change Log . . . . . . . . . . . . . . . . . . . . . 43 130 A.1. Changes between Version -00 and Version -01 of 131 draft-ietf-idnabis-rationale . . . . . . . . . . . . . . . 43 132 A.2. Version -02 . . . . . . . . . . . . . . . . . . . . . . . 44 133 A.3. Version -03 . . . . . . . . . . . . . . . . . . . . . . . 44 134 A.4. Version -04 . . . . . . . . . . . . . . . . . . . . . . . 44 135 A.5. Version -05 . . . . . . . . . . . . . . . . . . . . . . . 45 136 A.6. Version -06 . . . . . . . . . . . . . . . . . . . . . . . 45 137 A.7. Version -07 . . . . . . . . . . . . . . . . . . . . . . . 46 138 A.8. Version -08 . . . . . . . . . . . . . . . . . . . . . . . 46 139 A.9. Version -09 . . . . . . . . . . . . . . . . . . . . . . . 46 140 A.10. Version -10 . . . . . . . . . . . . . . . . . . . . . . . 47 141 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 47 143 1. Introduction 145 1.1. Context and Overview 147 Internationalized Domain Names in Applications (IDNA) is a collection 148 of standards that allow client applications to convert some Unicode 149 mnemonics to an ASCII-compatible encoding form ("ACE") which is a 150 valid DNS label containing only letters, digits, and hyphens. The 151 specific form of ACE label used by IDNA is called an "A-label". A 152 client can look up an exact A-label in the existing DNS, so A-labels 153 do not require any extensions to DNS, upgrades of DNS servers or 154 updates to low-level client libraries. An A-label is recognizable 155 from the prefix "xn--" before the characters produced by the Punycode 156 algorithm [RFC3492], thus a user application can identify an A-label 157 and convert it into Unicode (or some local coded character set) for 158 display. 160 [[anchor3: Note in draft: The above discussion, and the rest of the 161 text in this section, are very informal. In particular, the term 162 "A-label" is used to refer to some things that don't meet all of the 163 tests for A-labels. I have tightened it somewhat from the suggested 164 text I received, but not very much. Is the current form ok with 165 everyone???]] 167 On the registry side, IDNA allows a registry to offer 168 Internationalized Domain Names (IDNs) for registration as A-labels. 169 A registry may offer any subset of valid IDNs, and may apply any 170 restrictions or bundling (grouping of similar labels together in one 171 registration) appropriate for the context of that registry. 172 Registration of labels is sometimes discussed separately from lookup, 173 and is subject to a few specific requirements that do not apply to 174 lookup. 176 DNS clients and registries are subject to some differences in 177 requirements for handling IDNs. In particular, registries are urged 178 to register only exact, valid A-labels, while clients might do some 179 mapping to get from otherwise-invalid user input to a valid A-label. 181 The first version of IDNA was published in 2003 and is referred to 182 here as IDNA2003 to contrast it with the current version, which is 183 known as IDNA2008. The documents that made up both versions are 184 listed in Section 1.3.1. The characters that are valid in A-labels 185 are identified from rules listed in the Tables document 186 [IDNA2008-Tables], but validity can be derived from the Unicode 187 properties of those characters with a very few exceptions. 189 Traditionally, DNS labels are case-insensitive [RFC1034][RFC1035]. 190 That pattern was preserved in IDNA2003, but if case rules are 191 enforced from one language, another language sometimes loses the 192 ability to treat two characters separately. Case-sensitivity is 193 treated slightly differently in IDNA2008. 195 IDNA2003 used Unicode version 3.2 only. In order to keep up with new 196 characters added in new versions of UNICODE, IDNA2008 decouples its 197 rules from any particular version of UNICODE. Instead, the 198 attributes of new characters in Unicode determines how and whether 199 the characters can be used in IDNA labels. 201 This document provides informational context for IDNA2008, including 202 terminology, background, and policy discussions. 204 1.2. Discussion Forum 206 [[ RFC Editor: please remove this section. ]] 208 IDNA2008 is being discussed in the IETF "idnabis" Working Group and 209 on the mailing list idna-update@alvestrand.no 211 1.3. Terminology 213 Terminology for IDNA2008 appears in [IDNA2008-Defs]. That document 214 also contains a roadmap to the IDNA2008 document collection. No 215 attempt should be made to understand this document without the 216 definitions and concepts that appear there. 218 1.3.1. Documents and Standards 220 This document uses the term "IDNA2003" to refer to the set of 221 standards published in 2003 to define IDNA: the IDNA base 222 specification [RFC3490], Nameprep [RFC3491], Punycode [RFC3492], and 223 Stringprep [RFC3454]. 225 The term "IDNA2008" is used to refer to a new version of IDNA. 226 IDNA2008 is not dependent on any of the IDNA2003 specifications other 227 than the one for Punycode encoding. References to "these 228 specifications" or "these documents" are to the entire IDNA2008 set 229 listed in [IDNA2008-Defs]. 231 1.3.2. DNS "Name" Terminology 233 In the context of IDNs, the DNS term 'name' has introduced some 234 confusion as people speak of DNS labels in terms of the words or 235 phrases of various natural languages. Historically, many of the 236 "names" in the DNS have been mnemonics to identify some particular 237 concept, object, or organization. They are typically rooted in some 238 language because most people think in language-based ways. But, 239 because they are mnemonics, they need not obey the orthographic 240 conventions of any language: it is not a requirement that it be 241 possible for them to be "words". 243 This distinction is important because the reasonable goal of an IDN 244 effort is not to be able to write the great Klingon (or language of 245 one's choice) novel in DNS labels but to be able to form a usefully 246 broad range of mnemonics in ways that are as natural as possible in a 247 very broad range of scripts. 249 1.3.3. New Terminology and Restrictions 251 These documents introduce new terminology, and precise definitions, 252 for the terms "U-label", "A-Label", LDH-label (to which all valid 253 pre-IDNA host names conformed), Reserved-LDH-label (R-LDH-label), XN- 254 label, Fake-A-Label, and Non-Reserved-LDH-label (NR-LDH-label). 256 In addition, the term "putative label" has been adopted to refer to a 257 label that may appear to meet certain definitional constraints but 258 has not yet been sufficiently tested for validity. 260 These definitions are illustrated in Figure 1 of the Definitions 261 Document [IDNA2008-Defs]. R-LDH-labels contain "--" in the third and 262 fourth character from the beginning of the label. In IDNA-aware 263 applications, only a subset of these reserved labels is permitted to 264 be used, namely the A-label subset. A-labels are a subset of the 265 R-LDH-labels that begin with the case-insensitive string "xn--". 266 Labels that bear this prefix but which are not otherwise valid fall 267 into the "Fake-A-label" category. The non-reserved labels (NR-LDH- 268 labels) are implicitly valid since they do not trigger any 269 resemblance to IDNA-landr NR-LDH-labels. 271 The creation of the Reserved-LDH category is required for three 272 reasons: 274 o to prevent confusion with pre-IDNA coding forms; 276 o to permit future extensions that would require changing the 277 prefix, no matter how unlikely those might be (see Section 7.4); 278 and 280 o to reduce the opportunities for attacks via the Punycode encoding 281 algorithm itself. 283 1.4. Objectives 285 These are the main objectives in revising IDNA. 287 o Use a more recent version of Unicode, and allow IDNA to be 288 independent of Unicode versions, so that IDNA2008 need not be 289 update for implementations to adopt codepoints from new Unicode 290 versions. 292 o Fix a very small number of code-point categorizations that have 293 turned out to cause problems in the communities that use those 294 code-points. 296 o Reduce the dependency on mapping, in order that the pre-mapped 297 forms (which are not valid IDNA labels) tend to appear less often 298 in various contexts, in favor of valid A-labels. 300 o Fix some details in the bidirectional codepoint handling 301 algorithms. 303 1.5. Applicability and Function of IDNA 305 The IDNA specification solves the problem of extending the repertoire 306 of characters that can be used in domain names to include a large 307 subset of the Unicode repertoire. 309 IDNA does not extend DNS. Instead, the applications (and, by 310 implication, the users) continue to see an exact-match lookup 311 service. Either there is a single exactly-matching (subject to the 312 base DNS requirement of case-insensitive ASCII matching) name or 313 there is no match. This model has served the existing applications 314 well, but it requires, with or without internationalized domain 315 names, that users know the exact spelling of the domain names that 316 are to be typed into applications such as web browsers and mail user 317 agents. The introduction of the larger repertoire of characters 318 potentially makes the set of misspellings larger, especially given 319 that in some cases the same appearance, for example on a business 320 card, might visually match several Unicode code points or several 321 sequences of code points. 323 The IDNA standard does not require any applications to conform to it, 324 nor does it retroactively change those applications. An application 325 can elect to use IDNA in order to support IDN while maintaining 326 interoperability with existing infrastructure. If an application 327 wants to use non-ASCII characters in domain names, IDNA is the only 328 currently-defined option. Adding IDNA support to an existing 329 application entails changes to the application only, and leaves room 330 for flexibility in front-end processing and more specifically in the 331 user interface (see Section 6). 333 A great deal of the discussion of IDN solutions has focused on 334 transition issues and how IDNs will work in a world where not all of 335 the components have been updated. Proposals that were not chosen by 336 the original IDN Working Group would have depended on updating of 337 user applications, DNS resolvers, and DNS servers in order for a user 338 to apply an internationalized domain name in any form or coding 339 acceptable under that method. While processing must be performed 340 prior to or after access to the DNS, IDNA requires no changes to the 341 DNS protocol or any DNS servers or the resolvers on user's computers. 343 IDNA allows the graceful introduction of IDNs not only by avoiding 344 upgrades to existing infrastructure (such as DNS servers and mail 345 transport agents), but also by allowing some rudimentary use of IDNs 346 in applications by using the ASCII-encoded representation of the 347 labels containing non-ASCII characters. While such names are user- 348 unfriendly to read and type, and hence not optimal for user input, 349 they can be used as a last resort to allow rudimentary IDN usage. 350 For example, they might be the best choice for display if it were 351 known that relevant fonts were not available on the user's computer. 352 In order to allow user-friendly input and output of the IDNs and 353 acceptance of some characters as equivalent to those to be processed 354 according to the protocol, the applications need to be modified to 355 conform to this specification. 357 This version of IDNA uses the Unicode character repertoire, for 358 continuity with the original version of IDNA. 360 1.6. Comprehensibility of IDNA Mechanisms and Processing 362 One goal of IDNA2008, which is aided by the main goal of reducing the 363 dependency on mapping, is to improve the general understanding of how 364 to users and registrants are important design goals for this effort. 365 End-user applications have an important role to play in increasing 366 this comprehensibility. 368 Any system that tries to handle international characters encounters 369 some common problems. For example, a UI cannot display a character 370 if no font for that character is available. In some cases, 371 internationalization enables effective localization while maintaining 372 some global uniformity but losing some universality. 374 It is difficult to even make suggestions for end-user applications to 375 cope when characters and fonts are not available. Because display 376 functions are rarely controlled by the types of applications that 377 would call upon IDNA, such suggestions will rarely be very effective. 379 Converting between local character sets and normalized Unicode, if 380 needed, is part of this set of user agent issues. This conversion 381 introduces complexity in a system that is not Unicode-native. If a 382 label is converted to a local character set that does not have all 383 the needed characters, the user agent may have to add special logic 384 to avoid or reduce loss of information. 386 The major difficulty may lie in accurately identifying the incoming 387 character set and applying the correct conversion routine. Even more 388 difficult, the local character coding system could be based on 389 conceptually different assumptions than those used by Unicode (e.g., 390 choice of font encodings used for publications in some Indic 391 scripts). Those differences may not easily yield unambiguous 392 conversions or interpretations even if each coding system is 393 internally consistent and adequate to represent the local language 394 and script. 396 IDNA2008 shifts responsibility for character mapping and other 397 adjustments from the protocol (where it was located in IDNA2003) to 398 pre-processing before invoking IDNA. The intent is that this change 399 leads to greater usage of fully-valid A-Labels in display, transit 400 and storage, which should aid comprehensibility. A careful look at 401 pre-processing raises issues about what that pre-processing should do 402 and at what point pre-processing becomes harmful, how universally 403 consistent pre-processing algorithms can be, and how to be compatible 404 with labels prepared in a IDNA2003 context. Those issues are 405 discussed in Section 6. [[anchor9: Fix section reference.]] 407 2. Processing in IDNA2008 409 These specifications separate Domain Name Registration and Lookup in 410 the protocol specification. This separation reflects current 411 practice in which per-registry restrictions and special processing 412 are applied at registration time but not during lookup. Another 413 significant benefit is that separation facilitates incremental 414 addition of permitted character groups to avoid freezing on one 415 particular version of Unicode. 417 The actual registration and lookup protocols for IDNA2008 are 418 specified in [IDNA2008-Protocol]. 420 3. Permitted Characters: An Inclusion List 422 IDNA2008 adopts the inclusion model. A code-point is assumed to be 423 invalid, unless it is included as part of a Unicode property-based 424 rule or in rare cases included individually by an exception. When an 425 implementation moves to a new version of Unicode, the rules may 426 indicate new valid code-points. 428 This section provides an overview of the model used to establish the 429 algorithm and character lists of [IDNA2008-Tables] and describes the 430 names and applicability of the categories used there. Note that the 431 inclusion of a character in the first category group (Section 3.1.1) 432 does not imply that it can be used indiscriminately; some characters 433 are associated with contextual rules that must be applied as well. 435 The information given in this section is provided to make the rules, 436 tables, and protocol easier to understand. The normative generating 437 rules that correspond to this informal discussion appear in 438 [IDNA2008-Tables] and the rules that actually determine what labels 439 can be registered or looked up are in [IDNA2008-Protocol]. 441 3.1. A Tiered Model of Permitted Characters and Labels 443 Moving to an inclusion model involves a new specification for the 444 list of characters that are permitted in IDNs. In IDNA2003, 445 character validity is independent of context and fixed forever (or 446 until the standard is replaced). However, globally context- 447 independent rules have proved to be impractical because some 448 characters, especially those that are called "Join_Controls" in 449 Unicode, are needed to make reasonable use of some scripts but have 450 no visible effect in others. IDNA2003 prohibited those types of 451 characters entirely by discarding them. We now have a consensus that 452 under some conditions, these "joiner" characters are legitimately 453 needed to allow useful mnemonics for some languages and scripts. In 454 general, context-dependent rules help deal with characters that are 455 used differently across different scripts, and allow the standard to 456 be applied more appropriately in cases where a string is not 457 universally handled the same way. 459 IDNA2008 divides all possible Unicode code-points into four 460 categories: PROTOCOL-VALID, CONTEXTUAL RULE REQUIRED, DISALLOWED and 461 UNASSIGNED. 463 3.1.1. PROTOCOL-VALID 465 Characters identified as "PROTOCOL-VALID" (often abbreviated 466 "PVALID") are permitted in IDNs. Their use may be restricted by 467 rules about the context in which they appear or by other rules that 468 apply to the entire label in which they are to be embedded. For 469 example, any label that contains a character in this category that 470 has a "right-to-left" property must be used in context with the 471 "Bidi" rules (see [IDNA2008-Bidi]). 473 The term "PROTOCOL-VALID" is used to stress the fact that the 474 presence of a character in this category does not imply that a given 475 registry need accept registrations containing any of the characters 476 in the category. Registries are still expected to apply judgment 477 about labels they will accept and to maintain rules consistent with 478 those judgments (see [IDNA2008-Protocol] and Section 3.3). 480 Characters that are placed in the "PROTOCOL-VALID" category are 481 expected to never be removed from it or reclassified. While 482 theoretically characters could be removed from Unicode, such removal 483 would be inconsistent with the Unicode stability principles (see 484 [Unicode51], Appendix F) and hence should never occur. 486 3.1.2. CONTEXTUAL RULE REQUIRED 488 Some characters may be unsuitable for general use in IDNs but 489 necessary for the plausible support of some scripts. The two most 490 commonly-cited examples are the zero-width joiner and non-joiner 491 characters (ZWJ, U+200D and ZWNJ, U+200C). 493 3.1.2.1. Contextual Restrictions 495 Characters with contextual restrictions are identified as "CONTEXTUAL 496 RULE REQUIRED" and associated with a rule. The rule defines whether 497 the character is valid in a particular string, and also whether the 498 rule itself is to be applied on lookup as well as registration. 500 A distinction is made between characters that indicate or prohibit 501 joining and ones similar to them (known as "CONTEXT-JOINER" or 502 "CONTEXTJ") and other characters requiring contextual treatment 503 ("CONTEXT-OTHER" or "CONTEXTO"). Only the former require full 504 testing at lookup time. 506 It is important to note that these contextual rules cannot prevent 507 all uses of the relevant characters that might be confusing or 508 problematic. What they are expected do is to confine applicability 509 of the characters to scripts (and narrower contexts) where zone 510 administrators are knowledgeable enough about the use of those 511 characters to be prepared to deal with them appropriately. For 512 example, a registry dealing with an Indic script that requires ZWJ 513 and/or ZWNJ as part of the writing system is expected to understand 514 where the characters have visible effect and where they do not and to 515 make registration rules accordingly. By contrast, a registry dealing 516 with Latin or Cyrillic script might not be actively aware that the 517 characters exist, much less about the consequences of embedding them 518 in labels drawn from those scripts. 520 3.1.2.2. Rules and Their Application 522 Rules have descriptions such as "Must follow a character from Script 523 XYZ", "Must occur only if the entire label is in Script ABC", or 524 "Must occur only if the previous and subsequent characters have the 525 DFG property". The actual rules may be DEFINED or NULL. If present, 526 they may have values of "True" (character may be used in any position 527 in any label), "False" (character may not be used in any label), or 528 may be a set of procedural rules that specify the context in which 529 the character is permitted. 531 Examples of descriptions of typical rules, stated informally and in 532 English, include "Must follow a character from Script XYZ", "Must 533 occur only if the entire label is in Script ABC", "Must occur only if 534 the previous and subsequent characters have the DFG property". 536 Because it is easier to identify these characters than to know that 537 they are actually needed in IDNs or how to establish exactly the 538 right rules for each one, a rule may have a null value in a given 539 version of the tables. Characters associated with null rules are not 540 permitted to appear in putative labels for either registration or 541 lookup. Of course, a later version of the tables might contain a 542 non-null rule. 544 The actual rules and their descriptions are in [IDNA2008-Tables]. 545 [[anchor12: ??? Section number would be good here.]] That document 546 also creates a registry for future rules. 548 3.1.3. DISALLOWED 550 Some characters are inappropriate for use in IDNs and are thus 551 excluded for both registration and lookup (i.e., IDNA-conforming 552 applications performing name lookup should verify that these 553 characters are absent; if they are present, the label strings should 554 be rejected rather than converted to A-labels and looked up. Some of 555 these characters are problematic for use in IDNs (such as the 556 FRACTION SLASH character, U+2044), while some of them (such as the 557 various HEART symbols, e.g., U+2665, U+2661, and U+2765, see 558 Section 7.6) simply fall outside the conventions for typical 559 identifiers (basically letters and numbers). 561 Of course, this category would include code points that had been 562 removed entirely from Unicode should such removals ever occur. 564 Characters that are placed in the "DISALLOWED" category are expected 565 to never be removed from it or reclassified. If a character is 566 classified as "DISALLOWED" in error and the error is sufficiently 567 problematic, the only recourse would be either to introduce a new 568 code point into Unicode and classify it as "PROTOCOL-VALID" or for 569 the IETF to accept the considerable costs of an incompatible change 570 and replace the relevant RFC with one containing appropriate 571 exceptions. 573 There is provision for exception cases but, in general, characters 574 are placed into "DISALLOWED" if they fall into one or more of the 575 following groups: 577 o The character is a compatibility equivalent for another character. 578 In slightly more precise Unicode terms, application of 579 normalization method NFKC to the character yields some other 580 character. 582 o The character is an upper-case form or some other form that is 583 mapped to another character by Unicode casefolding. 585 o The character is a symbol or punctuation form or, more generally, 586 something that is not a letter, digit, or a mark that is used to 587 form a letter or digit. 589 3.1.4. UNASSIGNED 591 For convenience in processing and table-building, code points that do 592 not have assigned values in a given version of Unicode are treated as 593 belonging to a special UNASSIGNED category. Such code points are 594 prohibited in labels to be registered or looked up. The category 595 differs from DISALLOWED in that code points are moved out of it by 596 the simple expedient of being assigned in a later version of Unicode 597 (at which point, they are classified into one of the other categories 598 as appropriate). 600 The rationale for restricting the processing of UNASSIGNED characters 601 is simply that if such characters were permitted to be looked up, for 602 example, and were later assigned, but subject to some set of 603 contextual rules, un-updated instances of IDNA-aware software might 604 permit lookup of labels containing the previously-unassigned 605 characters while updated versions of IDNA-aware software might 606 restrict their use in lookup, depending on the contextual rules. It 607 should be clear that under no circumstance should an UNASSIGNED 608 character be permitted in a label to be registered as part of a 609 domain name. 611 3.2. Registration Policy 613 While these recommendations cannot and should not define registry 614 policies, registries should develop and apply additional restrictions 615 as needed to reduce confusion and other problems. For example, it is 616 generally believed that labels containing characters from more than 617 one script are a bad practice although there may be some important 618 exceptions to that principle. Some registries may choose to restrict 619 registrations to characters drawn from a very small number of 620 scripts. For many scripts, the use of variant techniques such as 621 those as described in RFC 3843 [RFC3743] and RFC 4290 [RFC4290], and 622 illustrated for Chinese by the tables described in RFC 4713 [RFC4713] 623 may be helpful in reducing problems that might be perceived by users. 625 In general, users will benefit if registries only permit characters 626 from scripts that are well-understood by the registry or its 627 advisers. If a registry decides to reduce opportunities for 628 confusion by constructing policies that disallow characters used in 629 historic writing systems or characters whose use is restricted to 630 specialized, highly technical contexts, some relevant information may 631 be found in Section 2.4 "Specific Character Adjustments", Table 4 632 "Candidate Characters for Exclusion from Identifiers" of 633 [Unicode-UAX31] and Section 3.1. "General Security Profile for 634 Identifiers" in [Unicode-Security]. 636 It is worth stressing that these principles of policy development and 637 application apply at all levels of the DNS, not only, e.g., TLD or 638 SLD registrations and that even a trivial, "anything permitted that 639 is valid under the protocol" policy is helpful in that it helps users 640 and application developers know what to expect. 642 3.3. Layered Restrictions: Tables, Context, Registration, Applications 644 The character rules in IDNA2008 are based on the realization that 645 there is no single magic bullet for any of the issues associated with 646 IDNs. Instead, the specifications define a variety of approaches. 647 The character tables are the first mechanism, protocol rules about 648 how those characters are applied or restricted in context are the 649 second, and those two in combination constitute the limits of what 650 can be done in the protocol. As discussed in the previous section 651 (Section 3.2), registries are expected to restrict what they permit 652 to be registered, devising and using rules that are designed to 653 optimize the balance between confusion and risk on the one hand and 654 maximum expressiveness in mnemonics on the other. 656 In addition, there is an important role for user agents in warning 657 against label forms that appear problematic given their knowledge of 658 local contexts and conventions. Of course, no approach based on 659 naming or identifiers alone can protect against all threats. 661 4. Issues that Constrain Possible Solutions 663 4.1. Display and Network Order 665 Domain names are always transmitted in network order (the order in 666 which the code points are sent in protocols), but may have a 667 different display order (the order in which the code points are 668 displayed on a screen or paper). When a domain name contains 669 characters that are normally written right to left, display order may 670 be affected although network order is not. It gets even more 671 complicated if left to right and right to left labels are adjacent to 672 each other within a domain name. The decision about the display 673 order is ultimately under the control of user agents --including Web 674 browsers, mail clients, hosted Web applications and many more -- 675 which may be highly localized. Should a domain name abc.def, in 676 which both labels are represented in scripts that are written right 677 to left, be displayed as fed.cba or cba.fed? Applications that are 678 in deployment today are already diverse, and one can find examples of 679 either choice. 681 The picture changes once again when an IDN appears in a 682 Internationalized Resource Identifier (IRI) [RFC3987]. An IRI or 683 Internationalized Email address contains elements other than the 684 domain name. For example, IRIs contain protocol identifiers and 685 field delimiter syntax such as "http://" or "mailto:" while email 686 addresses contain the "@" to separate local parts from domain names. 687 An IRI in network order begins with "http://" followed by domain 688 labels in network order, thus "http://abc.def". 690 User agents are not required to display and allow input of IRIs 691 directly but often do so. Implementors have to choose whether the 692 overall direction of these strings will always be left to right (or 693 right to left) for an IRI or email address. The natural order for a 694 user typing a domain name on a right to left system is fed.cba. 695 Should the R2L user agent reverse the entire domain name each time a 696 domain name is typed? Does this change if the user types "http://" 697 right before typing a domain name, thus implying that the user is 698 beginning at the beginning of the network order IRI? Experience in 699 the 1980s and 1990s with mixing systems in which domain name labels 700 were read in network order (left to right) and those in which those 701 labels were read right to left would predict a great deal of 702 confusion. 704 If each implementation of each application makes its own decisions on 705 these issues, users will develop heuristics that will sometimes fail 706 when switching applications. However, while some display order 707 conventions, voluntarily adopted, would be desirable to reduce 708 confusion, such suggestions are beyond the scope of these 709 specifications. 711 4.2. Entry and Display in Applications 713 Applications can accept and display domain names using any character 714 set or character coding system. That is, the IDNA protocol does not 715 necessarily affect the interface between users and applications. An 716 IDNA-aware application can accept and display internationalized 717 domain names in two formats: the internationalized character set(s) 718 supported by the application (i.e., an appropriate local 719 representation of a U-label), and as an A-label. Applications may 720 allow the display of A-labels, but are encouraged to not do so except 721 as an interface for special purposes, possibly for debugging, or to 722 cope with display limitations. In general, they should allow, but 723 not encourage, user input of A-labels. A-labels are opaque and ugly 724 and malicious variations on them are not easily detected by users. 725 Where possible, they should thus only be exposed when they are 726 absolutely needed. Because IDN labels can be rendered either as 727 A-labels or U-labels, the application may reasonably have an option 728 for the user to select the preferred method of display. Rendering 729 the U-label should normally be the default. 731 Domain names are often stored and transported in many places. For 732 example, they are part of documents such as mail messages and web 733 pages. They are transported in many parts of many protocols, such as 734 both the control commands of SMTP and associated the message body 735 parts, and in the headers and the body content in HTTP. It is 736 important to remember that domain names appear both in domain name 737 slots and in the content that is passed over protocols. 739 In protocols and document formats that define how to handle 740 specification or negotiation of charsets, labels can be encoded in 741 any charset allowed by the protocol or document format. If a 742 protocol or document format only allows one charset, the labels must 743 be given in that charset. Of course, not all charsets can properly 744 represent all labels. If a U-label cannot be displayed in its 745 entirety, the only choice (without loss of information) may be to 746 display the A-label. 748 Where a protocol or document format allows IDNs, labels should be in 749 whatever character encoding and escape mechanism the protocol or 750 document format uses at that place. This provision is intended to 751 prevent situations in which, e.g., UTF-8 domain names appear embedded 752 in text that is otherwise in some other character coding. 754 All protocols that use domain name slots (See Section 2.3.1.6 755 [[anchor15: ?? Verify this]] in [IDNA2008-Defs]) already have the 756 capacity for handling domain names in the ASCII charset. Thus, 757 A-labels can inherently be handled by those protocols. 759 4.3. Linguistic Expectations: Ligatures, Digraphs, and Alternate 760 Character Forms 762 Users have expectations about character matching or equivalence that 763 are based on their own languages and the orthography of those 764 languages. These expectations may not always be met in a global 765 system, especially if multiple languages are written using the same 766 script but using different conventions. Some examples: 768 o A Norwegian user might expect a label with the ae-ligature to be 769 treated as the same label as one using the Swedish spelling with 770 a-diaeresis even though applying that mapping to English would be 771 astonishing to users. 773 o A user in German might expect a label with an o-umlaut and a label 774 that had "oe" substituted, but was otherwise the same, treated as 775 equivalent even though that substitution would be a clear error in 776 Swedish. 778 o A Chinese user might expect automatic matching of Simplified and 779 Traditional Chinese characters, but applying that matching for 780 Korean or Japanese text would create considerable confusion. 782 o An English user might expect "theater" and "theatre" to match. 784 A number of languages use alphabetic scripts in which single phonemes 785 are written using two characters, termed a "digraph", for example, 786 the "ph" in "pharmacy" and "telephone". (Such characters can also 787 appear consecutively without forming a digraph, as in "tophat".) 788 Certain digraphs may be indicated typographically by setting the two 789 characters closer together than they would be if used consecutively 790 to represent different phonemes. Some digraphs are fully joined as 791 ligatures. For example, the word "encyclopaedia" is sometimes set 792 with a U+00E6 LATIN SMALL LIGATURE AE. When ligature and digraph 793 forms have the same interpretation across all languages that use a 794 given script, application of Unicode normalization generally resolves 795 the differences and causes them to match. When they have different 796 interpretations, matching must utilize other methods, presumably 797 chosen at the registry completely optional typographic convenience 798 for representing a digraph in one language (as in the above example 799 with some spelling conventions), while in another language it is a 800 single character that may not always be correctly representable by a 801 two-letter sequence (as in the above example with different spelling 802 conventions). This can be illustrated by many words in the Norwegian 803 language, where the "ae" ligature is the 27th letter of a 29-letter 804 extended Latin alphabet. It is equivalent to the 28th letter of the 805 Swedish alphabet (also containing 29 letters), U+00E4 LATIN SMALL 806 LETTER A WITH DIAERESIS, for which an "ae" cannot be substituted 807 according to current orthographic standards. 809 That character (U+00E4) is also part of the German alphabet where, 810 unlike in the Nordic languages, the two-character sequence "ae" is 811 usually treated as a fully acceptable alternate orthography for the 812 "umlauted a" character. The inverse is however not true, and those 813 two characters cannot necessarily be combined into an "umlauted a". 814 This also applies to another German character, the "umlauted o" 815 (U+00F6 LATIN SMALL LETTER O WITH DIAERESIS) which, for example, 816 cannot be used for writing the name of the author "Goethe". It is 817 also a letter in the Swedish alphabet where, like the "a with 818 diaeresis", it cannot be correctly represented as "oe" and in the 819 Norwegian alphabet, where it is represented, not as "o with 820 diaeresis", but as "slashed o", U+00F8. 822 Some of the ligatures that have explicit code points in Unicode were 823 given special handling in IDNA2003 and now pose additional problems 824 in transition. See Section 7.2. 826 Additional cases with alphabets written right to left are described 827 in Section 4.5. 829 Matching and comparison algorithm selection often requires 830 information about the language being used, context, or both -- 831 information that is not available to IDNA or the DNS. Consequently, 832 these specifications make no attempt to treat combined characters in 833 any special way. A registry that is aware of the language context in 834 which labels are to be registered, and where that language sometimes 835 (or always) treats the two- character sequences as equivalent to the 836 combined form, should give serious consideration to applying a 837 "variant" model [RFC3743] [RFC4290], or to prohibiting registration 838 of one the forms entirely, to reduce the opportunities for user 839 confusion and fraud that would result from the related strings being 840 registered to different parties. 842 [[anchor16: Placeholder: A discussion of the Arabic digit issue 843 should go here once it is resolved in some appropriate way.]] 845 4.4. Case Mapping and Related Issues 847 In the DNS, ASCII letters are stored with their case preserved. 848 Matching during the query process is case-independent, but none of 849 the information that might be represented by choices of case has been 850 lost. That model has been accidentally helpful because, as people 851 have created DNS labels by catenating words (or parts of words) to 852 form labels, case has often been used to distinguish among components 853 and make the labels more memorable. 855 Since DNS servers do not get involved in parsing IDNs, they cannot do 856 case-independent matching. Thus, keeping the cases separate in 857 lookup or registration, and doing matching at the server, is not 858 feasible with IDNA or any similar approach. Case-matching must be 859 done, if desired, by IDN clients even though it wasn't done by ASCII- 860 only DNS clients. That situation was recognized in IDNA2003 and 861 nothing in these specifications fundamentally changes it or could do 862 so. In IDNA2003, all characters are case-folded and mapped by 863 clients in a standardized step. 865 Some characters do not have upper case forms. For example the 866 Unicode case folding operation maps Greek Final Form Sigma (U+03C2) 867 to the medial form (U+03C3) and maps Eszett (German Sharp S, U+00DF) 868 to "ss". Neither of these mappings is reversible because the upper 869 case of U+03C3 is the Upper Case Sigma (U+03A3) and "ss" is an ASCII 870 string. IDNA2008 permits, at the risk of some incompatibility, 871 slightly more flexibility in this area by avoid case folding and 872 treating these characters as themselves. Approaches to handling one- 873 way mappings are discussed in Section 7.2. 875 Because IDNA2003 maps Final Sigma and Eszett to other characters, and 876 the reverse mapping is never possible, that in some sense means that 877 neither Final Sigma nor Eszett can be represented in a IDNA2003 IDN. 878 With IDNA2008, both characters can be used in an IDN and so the 879 A-label used for lookup for any U-label containing those characters, 880 is now different. See Section 7.1 for a discussion of what kinds of 881 changes might require the IDNA prefix to change; this case is clearly 882 worth discussing but the WG came to consensus not to make a prefix 883 change anyway. 885 4.5. Right to Left Text 887 In order to be sure that the directionality of right to left text is 888 unambiguous, IDNA2003 required that any label in which right to left 889 characters appear both starts and ends with them and that it not 890 include any characters with strong left to right properties (that 891 excludes other alphabetic characters but permits European digits). 892 Any other string that contains a right to left character and does not 893 meet those requirements is rejected. This is one of the few places 894 where the IDNA algorithms (both in IDNA2003 and in IDAN2008) examine 895 an entire label, not just individual characters. The algorithmic 896 model used in IDNA2003 rejects the label when the final character in 897 a right to left string requires a combining mark in order to be 898 correctly represented. 900 That prohibition is not acceptable for writing systems for languages 901 written with consonantal alphabets to which diacritical vocalic 902 systems are applied, and for languages with orthographies derived 903 from them where the combining marks may have different functionality. 904 In both cases the combining marks can be essential components of the 905 orthography. Examples of this are Yiddish, written with an extended 906 Hebrew script, and Dhivehi (the official language of Maldives) which 907 is written in the Thaana script (which is, in turn, derived from the 908 Arabic script). IDNA2008 removes the restriction on final combining 909 characters with a new set of rules for right to left scripts and 910 their characters. Those new rules are specified in [IDNA2008-Bidi]. 912 5. IDNs and the Robustness Principle 914 The "Robustness Principle" is often stated as "Be conservative about 915 what you send and liberal in what you accept" (See, e.g., Section 916 1.2.2 of the applications-layer Host Requirements specification 917 [RFC1123]) This principle applies to IDNA. In applying the principle 918 to registries as the source ("sender") of all registered and useful 919 IDNs, registries are responsible for being conservative about what 920 they register and put out in the Internet. For IDNs to work well, 921 zone administrators (registries) must have and require sensible 922 policies about what is registered -- conservative policies -- and 923 implement and enforce them. 925 Conversely, lookup applications are expected to reject labels that 926 clearly violate global (protocol) rules (no one has ever seriously 927 claimed that being liberal in what is accepted requires being 928 stupid). However, once one gets past such global rules and deals 929 with anything sensitive to script or locale, it is necessary to 930 assume that garbage has not been placed into the DNS, i.e., one must 931 be liberal about what one is willing to look up in the DNS rather 932 than guessing about whether it should have been permitted to be 933 registered. 935 If a string cannot be successfully found in the DNS after the lookup 936 processing described here, it makes no difference whether it simply 937 wasn't registered or was prohibited by some rule at the registry. 938 Application implementors should be aware that where DNS wildcards are 939 used, the ability to successfully resolve a name does not guarantee 940 that it was actually registered. 942 6. Front-end and User Interface Processing for Lookup 944 [[anchor18: Note in Draft: While this section has been revised in 945 version -10 to improve clarity, a significant revision is expected 946 once the discussions of mapping stabilize.]] 947 Domain names may be identified and processed in many contexts. They 948 may be typed in by users either by themselves or embedded in an 949 identifier such as email addresses, URIs, or IRIs. They may occur in 950 running text or be processed by one system after being provided in 951 another. Systems may try to normalize URLs to determine (or guess) 952 whether a reference is valid or two references point to the same 953 object without actually looking the objects up (comparison without 954 lookup is necessary for URI types that are not intended to be 955 resolved). Some of these goals may be more easily and reliably 956 satisfied than others. While there are strong arguments for any 957 domain name that is placed "on the wire" -- transmitted between 958 systems -- to be in the zero-ambiguity forms of A-labels, it is 959 inevitable that programs that process domain names will encounter 960 U-labels or variant forms. 962 This section discusses these mapping and transformation issues among 963 names, contrasting IDNA2003 and IDNA2008 behavior. The discussion 964 applies only in operations that look up names or interpret files. 965 There are several reasons why registration activities should require 966 final names and verification of those names by the would-be 967 registrant. 969 One source of label forms that are neither A-labels nor U-labels will 970 be labels created under IDNA2003. That protocol allowed labels that 971 were transformed from native-character format by mapping some 972 characters into others before conversion into A-label format. One 973 consequence of the transformations was that conversion from the 974 A-label format back to native characters often did not produce the 975 original label. IDNA2008 explicitly defines A-labels and U-labels as 976 different forms of the same abstract label, forms that are stable 977 when conversions are performed between them (without mappings). 979 A different way of explaining this is that there are, today, domain 980 names in files on the Internet that use characters that cannot be 981 represented directly in, or recovered from, (A-label) domain names 982 but for which interpretations were provided by IDNA2003). There are 983 two major categories of characters irreversibly remapped by 984 Stringprep, those that are removed by NFKC normalization and those 985 upper-case characters that are mapped to lower-case (there are also a 986 few characters that are given special-case mapping treatment, 987 including lower-case characters that are case-folded into other 988 lower-case characters or strings and those that are simply 989 eliminated). 991 Other issues in domain name identification and processing arise 992 because IDNA2003 specified that several other characters be treated 993 as equivalent to the ASCII period (dot, full stop) character used as 994 a label separator. If a string that might be a domain name appears 995 in an arbitrary context (such as running text), it is difficult, even 996 with only ASCII characters, to know whether an actual domain name (or 997 a protocol parameter like a URI) is present and where it starts and 998 ends. When using Unicode, this gets even more difficult if treatment 999 of certain special characters (like the dot that separates labels in 1000 a domain name) depends on context (e.g., prior knowledge of whether 1001 the string represents a domain name or not). That knowledge is not 1002 available if the primary heuristic for identifying the presence of 1003 domain names in strings depends on the presence of dots separating 1004 groups of characters with no intervening spaces. 1006 [[anchor19: Placeholder: In serial efforts to move the mapping model 1007 out of the protocol and leave it unspecified here, this paragraph has 1008 become a complete botch. Rewrite when the mapping plan stabilizes.]] 1009 The IDNA2008 model removes all of these mappings and interpretations, 1010 including the equivalence of different forms of dots, from the 1011 protocol, discouraging such mappings and leaving them, when 1012 necessary, to local processing. This should not be taken to imply 1013 that local processing is optional or can be avoided entirely, even if 1014 doing so might have been desirable in a world without IDNA2003 IDNs 1015 in files and archives. Instead, unless the program context is such 1016 that it is known that any IDNs that appear will contain either 1017 U-label or A-label forms, or that other forms can safely be rejected, 1018 some local processing of apparent domain name strings will be 1019 required, both to maintain compatibility with IDNA2003 and to prevent 1020 user astonishment. Such local processing, while not specified in 1021 this document or the associated ones, will generally take one of two 1022 forms: 1024 o Generic Preprocessing. 1025 When the context in which the program or system that processes 1026 domain names operates is global, a reasonable balance must be 1027 found that is sensitive to the broad range of local needs and 1028 assumptions while, at the same time, not sacrificing the needs of 1029 one language, script, or user population to those of another. 1031 For this case, the best practice will usually be to apply NFKC and 1032 case-mapping (or, perhaps better yet, Stringprep itself), plus 1033 dot-mapping where appropriate, to the domain name string prior to 1034 applying IDNA. That practice will not only yield a reasonable 1035 compromise of user experience with protocol requirements but will 1036 be almost completely compatible with the various forms permitted 1037 by IDNA2003. 1039 o Highly Localized Preprocessing. 1040 Unlike the case above, there will be some situations in which 1041 software will be highly localized for a particular environment and 1042 carefully adapted to the expectations of users in that 1043 environment. The many discussions about using the Internet to 1044 preserve and support local cultures suggest that these cases may 1045 be more common in the future than they have been so far. 1047 In these cases, we should avoid trying to tell implementers what 1048 they should accept, if only because they are quite likely (and for 1049 good reason) to ignore us. We would assume that they would map 1050 characters that the intuitions of their users would suggest be 1051 mapped and would hope that they would do that mapping as early as 1052 possible, storing A-label or U-label forms in files and 1053 transporting only those forms between systems. One can imagine 1054 switches about whether some sorts of mappings occur, warnings 1055 before applying them or, in a slightly more extreme version of the 1056 approach taken in Internet Explorer version 7 (IE7), systems that 1057 utterly refuse to handle "strange" characters at all if they 1058 appear in U-label form. None of those local decisions are a 1059 threat to interoperability as long as (i) only U-labels and 1060 A-labels are used in interchange with systems outside the local 1061 environment, (ii) no character that would be valid in a U-label as 1062 itself is mapped to something else, (iii) any local mappings are 1063 applied as a preprocessing step (or, for conversions from U-labels 1064 or A-labels to presentation forms, postprocessing), not as part of 1065 IDNA processing proper, and (iv) appropriate consideration is 1066 given to labels that might have entered the environment in 1067 conformance to IDNA2003. 1069 In either case, it is vital that user interface designs and, where 1070 the interfaces are not sufficient, users, be aware that the only 1071 forms of domain names that this protocol anticipates will resolve 1072 globally or compare equal when crude methods (i.e., those not 1073 conforming to the strict definition of label equivalence given in 1074 [IDNA2008-Defs]) are used are those in which all native-script labels 1075 are in U-label form. Forms that assume mapping will occur, 1076 especially forms that were not valid under IDNA2003, may or may not 1077 function in predictable ways across all implementations. 1079 User interfaces involving Latin-based scripts should take special 1080 care when considering how to handle case mapping because small 1081 differences in label strings may cause behavior that is astonishing 1082 to users. Because case-insensitive comparison is done for ASCII 1083 strings by DNS-servers, an all-ASCII label is treated as case- 1084 insensitive. However, if even one of the characters of that string 1085 is replaced by one that requires the label to be given IDN treatment 1086 (e.g., by adding a diacritical mark), then the label effectively 1087 becomes case-sensitive because only lower-case characters are 1088 permitted in IDNs. Preprocessing in applications to handle case 1089 mapping for Latin-based scripts (and possibly other scripts with case 1090 distinctions) may be wise to prevent user astonishment. However, all 1091 applications may not do this and ambiguity in transport is not 1092 desirable. Consequently the case-dependent forms should not be 1093 stored in files. 1095 7. Migration from IDNA2003 and Unicode Version Synchronization 1097 7.1. Design Criteria 1099 As mentioned above and in RFC 4690, two key goals of the IDNA2008 1100 design are 1102 o to enable applications to be agnostic about whether they are being 1103 run in environments supporting any Unicode version from 3.2 1104 onward, 1106 o to permit incrementally adding new characters, character groups, 1107 scripts, and other character collections as they are incorporated 1108 into Unicode, doing so without disruption and, in the long term, 1109 without "heavy" processes (an IETF consensus process is required 1110 by the IDNA2008 specifications and is expected to be required and 1111 used until significant experience accumulates with IDNA operations 1112 and new versions of Unicode). 1114 7.1.1. Summary and Discussion of IDNA Validity Criteria 1116 The general criteria for a label to be considered IDNA-valid are (the 1117 actual rules are rigorously defined in the "Protocol" and "Tables" 1118 documents): 1120 o The characters are "letters", marks needed to form letters, 1121 numerals, or other code points used to write words in some 1122 language. Symbols, drawing characters, and various notational 1123 characters are intended to be permanently excluded. There is no 1124 evidence that they are important enough to Internet operations or 1125 internationalization to justify expansion of domain names beyond 1126 the general principle of "letters, digits, and hyphen". 1127 (Additional discussion and rationale for the symbol decision 1128 appears in Section 7.6). 1130 o Other than in very exceptional cases, e.g., where they are needed 1131 to write substantially any word of a given language, punctuation 1132 characters are excluded. The fact that a word exists is not proof 1133 that it should be usable in a DNS label and DNS labels are not 1134 expected to be usable for multiple-word phrases (although they are 1135 certainly not prohibited if the conventions and orthography of a 1136 particular language cause that to be possible). 1138 o Characters that are unassigned (have no character assignment at 1139 all) in the version of Unicode being used by the registry or 1140 application are not permitted, even on lookup. The issues 1141 involved in this decision are discussed in Section 7.7. 1143 o Any character that is mapped to another character by a current 1144 version of NFKC is prohibited as input to IDNA (for either 1145 registration or lookup). With a few exceptions, this principle 1146 excludes any character mapped to another by Nameprep [RFC3491]. 1148 The principles above drive the design of rules that are specified 1149 exactly in [IDNA2008-Tables]. Those rules identify the characters 1150 that are IDNA-valid. The rules themselves are normative, and the 1151 tables are derived from them, rather than vice versa. 1153 7.1.2. Labels in Registration 1155 Any label registered in a DNS zone must be validated -- i.e., the 1156 criteria for that label must be met -- in order for applications to 1157 work as intended. This principle is not new. For example, since the 1158 DNS was first deployed, zone administrators have been expected to 1159 verify that names meet "hostname" requirements [RFC0952] where those 1160 requirements are imposed by the expected applications. Other 1161 applications contexts, such as the later addition of special service 1162 location formats [RFC2782] imposed new requirements on zone 1163 administrators. For zones that will contain IDNs, support for 1164 Unicode version-independence requires restrictions on all strings 1165 placed in the zone. In particular, for such zones: 1167 o Any label that appears to be an A-label, i.e., any label that 1168 starts in "xn--", must be IDNA-valid, i.e., they must be valid 1169 A-labels, as discussed in Section 2 above. 1171 o The Unicode tables (i.e., tables of code points, character 1172 classes, and properties) and IDNA tables (i.e., tables of 1173 contextual rules such as those that appear in the Tables 1174 document), must be consistent on the systems performing or 1175 validating labels to be registered. Note that this does not 1176 require that tables reflect the latest version of Unicode, only 1177 that all tables used on a given system are consistent with each 1178 other. 1180 Under this model, registry tables will need to be updated (both the 1181 Unicode-associated tables and the tables of permitted IDN characters) 1182 to enable a new script or other set of new characters. The registry 1183 will not be affected by newer versions of Unicode, or newly- 1184 authorized characters, until and unless it wishes to support them. 1185 The zone administrator is responsible for verifying IDNA-validity as 1186 well as its local policies -- a more extensive set of checks than are 1187 required for looking up the labels. Systems looking up or resolving 1188 DNS labels, especially IDN DNS labels, must be able to assume that 1189 applicable registration rules were followed for names entered into 1190 the DNS. 1192 7.1.3. Labels in Lookup 1194 Anyone looking up a label in a DNS zone is required to 1196 o Maintain IDNA and Unicode tables that are consistent with regard 1197 to versions, i.e., unless the application actually executes the 1198 classification rules in [IDNA2008-Tables], its IDNA tables must be 1199 derived from the version of Unicode that is supported more 1200 generally on the system. As with registration, the tables need 1201 not reflect the latest version of Unicode but they must be 1202 consistent. 1204 o Validate the characters in labels to be looked up only to the 1205 extent of determining that the U-label does not contain 1206 "DISALLOWED" code points or code points that are unassigned in its 1207 version of Unicode. 1209 o Validate the label itself for conformance with a small number of 1210 whole-label rules. In particular, it must verify that 1212 * there are no leading combining marks, 1214 * the "bidi" conditions are met if right to left characters 1215 appear, 1217 * any required contextual rules are available, and 1219 * any contextual rules that are associated with Joiner Controls 1220 are tested. 1222 o Do not reject labels based on other contextual rules about 1223 characters, including mixed-script label prohibitions. Such rules 1224 may be used to influence presentation decisions in the user 1225 interface, but not to avoid looking up domain names. 1227 Lookup applications that following these rules, rather than having 1228 their own criteria for rejecting lookup attempts, are not sensitive 1229 to version incompatibilities with the particular zone registry 1230 associated with the domain name except for labels containing 1231 characters recently added to Unicode. 1233 An application or client that processes names according to this 1234 protocol and then resolves them in the DNS will be able to locate any 1235 name that is registered, as long as those registrations are IDNA- 1236 value and its version of the IDNA tables is sufficiently up-to-date 1237 to interpret all of the characters in the label. Messages to users 1238 should distinguish between "label contains an unallocated code point" 1239 and other types of lookup failures. A failure on the basis of an old 1240 version of Unicode may lead the user to a desire to upgrade to a 1241 newer version, but will have no other ill effects (this is consistent 1242 with behavior in the transition to the DNS when some hosts could not 1243 yet handle some forms of names or record types). 1245 7.2. Changes in Character Interpretations 1247 [[anchor22: This subsection will need to be rewritten when the 1248 mapping decisions stabilize.]] 1250 In those scripts that make case distinctions, there are a few 1251 characters for which an obvious and unique upper case character has 1252 not historically been available to match a lower case one or vice 1253 versa. For those characters, the mappings used in constructing the 1254 Stringprep tables for IDNA2003, performed using the Unicode CaseFold 1255 operation (See Section 5.8 of the Unicode Standard [Unicode51]), 1256 generate different characters or sets of characters. Those 1257 operations are not reversible and lose even more information than 1258 traditional upper case or lower case transformations, but are more 1259 useful than those transformations for comparison purposes. Two 1260 notable characters of this type are the German character Eszett 1261 (Sharp S, U+00DF) and the Greek Final Form Sigma (U+03C2). The 1262 former is case-folded to the ASCII string "ss", the latter to a 1263 medial (Lower Case) Sigma (U+03C3). 1265 The decision to eliminate mappings, including case folding, from the 1266 IDNA2008 protocol in order to make A-labels and U-labels idempotent 1267 made these characters problematic. If they were to be disallowed, 1268 important words and mnemonics could not be written in 1269 orthographically reasonable ways. If they were to be permitted as 1270 distinct characters, there would be no information loss and 1271 registries would have more flexibility, but IDNA2003 and IDNA2008 1272 lookups might result in different A-labels. 1274 With the understanding that there would be incompatibility either way 1275 but a judgment that the incompatibility was not significant enough to 1276 justify a prefix change, the WG concluded that Eszett and Final Form 1277 Sigma should be treated as distinct and Protocol-Valid characters. 1279 Registries, especially those maintaining zones for third parties, 1280 must decide how to introduce a new service in a way that does not 1281 create confusion or significantly weaken or invalidate existing 1282 identifiers. This is not a new problem; registries were faced with 1283 similar issues when IDNs were introduced and when other new forms of 1284 strings have been permitted as labels. 1286 There are several approaches to problems of this type. Without any 1287 preference or claim to completeness, some of these, all of which have 1288 been used by registries in the past for similar transitions, are: 1290 o Do not permit use of the newly-available character at the registry 1291 level. This might cause lookup failures if a domain name were to 1292 be written with the expectation of the IDNA2003 mapping behavior, 1293 but would eliminate any possibility of false matches. 1295 o Hold a "sunrise"-like arrangement in which holders of labels 1296 containing "ss" in the Eszett case or Lower Case Sigma are given 1297 priority (and perhaps other benefits) for registering the 1298 corresponding string containing Eszett or Final Sigma 1299 respectively. 1301 o Adopt some sort of "variant" approach in which registrants obtain 1302 labels with both character forms. 1304 o Adopt a different form of "variant" approach in which registration 1305 of additional names is either not permitted at all or permitted 1306 only by the registrant who already has one of the names. 1308 7.3. More Flexibility in User Agents 1310 [[anchor23: Note in Draft: This section is mapping-related and may 1311 need to be revised after that issue settles down.]] Also, it is 1312 closely related to Section 4.2 and may need to be cross-referenced 1313 from it or consolidated into it. 1315 These documents do not specify mappings between one character or code 1316 point and others. Instead, IDNA2008 prohibits characters that would 1317 be mapped to others by normalization, upper case to lower case 1318 changes, or other rules. As examples, while mathematical characters 1319 based on Latin ones are accepted as input to IDNA2003, they are 1320 prohibited in IDNA2008. Similarly, double-width characters and other 1321 variations are prohibited as IDNA input. 1323 Since the rules in [IDNA2008-Tables] have the effect that only 1324 strings that are not transformed by NFKC are valid, if an application 1325 chooses to perform NFKC normalization before lookup, that operation 1326 is safe since this will never make the application unable to look up 1327 any valid string. However, as discussed above, the application 1328 cannot guarantee that any other application will perform that 1329 mapping, so it should be used only with caution and for informed 1330 users. 1332 In many cases these prohibitions should have no effect on what the 1333 user can type as input to the lookup process. It is perfectly 1334 reasonable for systems that support user interfaces to perform some 1335 character mapping that is appropriate to the local environment. This 1336 would normally be done prior to actual invocation of IDNA. At least 1337 conceptually, the mapping would be part of the Unicode conversions 1338 discussed above and in [IDNA2008-Protocol]. However, those changes 1339 will be local ones only -- local to environments in which users will 1340 clearly understand that the character forms are equivalent. For use 1341 in interchange among systems, it appears to be much more important 1342 that U-labels and A-labels can be mapped back and forth without loss 1343 of information. 1345 One specific, and very important, instance of this strategy arises 1346 with case-folding. In the ASCII-only DNS, names are looked up and 1347 matched in a case-independent way, but no actual case-folding occurs. 1348 Names can be placed in the DNS in either upper or lower case form (or 1349 any mixture of them) and that form is preserved, returned in queries, 1350 and so on. IDNA2003 approximated that behavior for non-ASCII strings 1351 by performing case-folding at registration time (resulting in only 1352 lower-case IDNs in the DNS) and when names were looked up. 1354 As suggested earlier in this section, it appears to be desirable to 1355 do as little character mapping as possible as long as Unicode works 1356 correctly (e.g., NFC mapping to resolve different codings for the 1357 same character is still necessary although the specifications require 1358 that it be performed prior to invoking the protocol) in order to make 1359 the mapping between A-labels and U-labels idempotent. Case-mapping 1360 is not an exception to this principle. If only lower case characters 1361 can be registered in the DNS (i.e., be present in a U-label), then 1362 IDNA2008 should prohibit upper-case characters as input. Some other 1363 considerations reinforce this conclusion. For example, in ASCII 1364 case-mapping for individual characters, uppercase(character) must be 1365 equal to uppercase(lowercase(character)). That may not be true with 1366 IDNs. In some scripts that use case distinctions, there are a few 1367 characters that do not have counterparts in one case or the other. 1368 The relationship between upper case and lower case may even be 1369 language-dependent, with different languages (or even the same 1370 language in different areas) expecting different mappings. User 1371 agents can meet the expectations of users who are accustomed to the 1372 case-insensitive DNS environment by performing case folding prior to 1373 IDNA processing, but the IDNA procedures themselves should neither 1374 require such mapping nor expect them when they are not natural to the 1375 localized environment. 1377 7.4. The Question of Prefix Changes 1379 The conditions that would require a change in the IDNA ACE prefix 1380 ("xn--" for the version of IDNA specified in [RFC3490]) have been a 1381 great concern to the community. A prefix change would clearly be 1382 necessary if the algorithms were modified in a manner that would 1383 create serious ambiguities during subsequent transition in 1384 registrations. This section summarizes our conclusions about the 1385 conditions under which changes in prefix would be necessary and the 1386 implications of such a change. 1388 7.4.1. Conditions Requiring a Prefix Change 1390 An IDN prefix change is needed if a given string would be looked up 1391 or otherwise interpreted differently depending on the version of the 1392 protocol or tables being used. An IDNA upgrade would require a 1393 prefix change if, and only if, one of the following four conditions 1394 were met: 1396 1. The conversion of an A-label to Unicode (i.e., a U-label) yields 1397 one string under IDNA2003 (RFC3490) and a different string under 1398 IDNA2008. 1400 2. In a significant number of cases, an input string that is valid 1401 under IDNA2003 and also valid under IDNA2008 yields two different 1402 A-labels with the different versions. This condition is believed 1403 to be essentially equivalent to the one above except for a very 1404 small number of edge cases which may not justify a prefix change 1405 (See Section 7.2). 1407 Note that if the input string is valid under one version and not 1408 valid under the other, this condition does not apply. See the 1409 first item in Section 7.4.2, below. 1411 3. A fundamental change is made to the semantics of the string that 1412 is inserted in the DNS, e.g., if a decision were made to try to 1413 include language or script information in the encoding in 1414 addition to the string itself. 1416 4. A sufficiently large number of characters is added to Unicode so 1417 that the Punycode mechanism for block offsets can no longer 1418 reference the higher-numbered planes and blocks. This condition 1419 is unlikely even in the long term and certain not to arise in the 1420 next several years. 1422 7.4.2. Conditions Not Requiring a Prefix Change 1424 As a result of the principles described above, none of the following 1425 changes require a new prefix: 1427 1. Prohibition of some characters as input to IDNA. This may make 1428 names that are now registered inaccessible, but does not change 1429 those names. 1431 2. Adjustments in IDNA tables or actions, including normalization 1432 definitions, that affect characters that were already invalid 1433 under IDNA2003. 1435 3. Changes in the style of the IDNA definition that does not alter 1436 the actions performed by IDNA. 1438 7.4.3. Implications of Prefix Changes 1440 While it might be possible to make a prefix change, the costs of such 1441 a change are considerable. Registries could not convert all IDNA2003 1442 ("xn--") registrations to a new form at the same time and synchronize 1443 that change with applications supporting lookup. Unless all existing 1444 registrations were simply to be declared invalid (and perhaps even 1445 then) systems that needed to support both labels with old prefixes 1446 and labels with new ones would first process a putative label under 1447 the IDNA2008 rules and try to look it up and then, if it were not 1448 found, would process the label under IDNA2003 rules and look it up 1449 again. That process could significantly slow down all processing 1450 that involved IDNs in the DNS especially since a fully-qualified name 1451 might contain a mixture of labels that were registered with the old 1452 and new prefixes. That would make DNS caching very difficult. In 1453 addition, looking up the same input string as two separate A-labels 1454 creates some potential for confusion and attacks, since the labels 1455 could map to different targets and then resolve to different entries 1456 in the DNS. 1458 Consequently, a prefix change is to be avoided if at all possible, 1459 even if it means accepting some IDNA2003 decisions about character 1460 distinctions as irreversible and/or giving special treatment to edge 1461 cases. 1463 7.5. Stringprep Changes and Compatibility 1465 The Nameprep [RFC3491] specification, a key part of IDNA2003, is a 1466 profile of Stringprep [RFC3454]. While Nameprep is a Stringprep 1467 profile specific to IDNA, Stringprep is used by a number of other 1468 protocols. Were Stringprep to be modified by IDNA2008, those changes 1469 to improve the handling of IDNs could cause problems for non-DNS 1470 uses, most notably if they affected identification and authentication 1471 protocols. Several elements of IDNA2008 give interpretations to 1472 strings prohibited under IDNA2003 or prohibit strings that IDNA2003 1473 permitted. Those elements include the proposed new inclusion tables 1474 [IDNA2008-Tables], the reduction in the number of characters 1475 permitted as input for registration or lookup (Section 3), and even 1476 the proposed changes in handling of right to left strings 1477 [IDNA2008-Bidi]. IDNA2008 does not use Nameprep or Stringprep at 1478 all, so there are no side-effect changes to other protocols. 1480 It is particularly important to keep IDNA processing separate from 1481 processing for various security protocols because some of the 1482 constraints that are necessary for smooth and comprehensible use of 1483 IDNs may be unwanted or undesirable in other contexts. For example, 1484 the criteria for good passwords or passphrases are very different 1485 from those for desirable IDNs: passwords should be hard to guess, 1486 while domain names should normally be easily memorable. Similarly, 1487 internationalized SCSI identifiers and other protocol components are 1488 likely to have different requirements than IDNs. 1490 7.6. The Symbol Question 1492 One of the major differences between this specification and the 1493 original version of IDNA is that the original version permitted non- 1494 letter symbols of various sorts, including punctuation and line- 1495 drawing symbols, in the protocol. They were always discouraged in 1496 practice. In particular, both the "IESG Statement" about IDNA and 1497 all versions of the ICANN Guidelines specify that only language 1498 characters be used in labels. This specification disallows symbols 1499 entirely. There are several reasons for this, which include: 1501 1. As discussed elsewhere, the original IDNA specification assumed 1502 that as many Unicode characters as possible should be permitted, 1503 directly or via mapping to other characters, in IDNs. This 1504 specification operates on an inclusion model, extrapolating from 1505 the original "hostname" rules (LDH, see [IDNA2008-Defs]) -- which 1506 have served the Internet very well -- to a Unicode base rather 1507 than an ASCII base. 1509 2. Symbol names are more problematic than letters because there may 1510 be no general agreement on whether a particular glyph matches a 1511 symbol; there are no uniform conventions for naming; variations 1512 such as outline, solid, and shaded forms may or may not exist; 1513 and so on. As just one example, consider a "heart" symbol as it 1514 might appear in a logo that might be read as "I love...". While 1515 the user might read such a logo as "I love..." or "I heart...", 1516 considerable knowledge of the coding distinctions made in Unicode 1517 is needed to know that there more than one "heart" character 1518 (e.g., U+2665, U+2661, and U+2765) and how to describe it. These 1519 issues are of particular importance if strings are expected to be 1520 understood or transcribed by the listener after being read out 1521 loud. 1522 [[anchor24: The above paragraph remains controversial as to 1523 whether it is valid. The WG will need to make a decision if this 1524 section is not dropped entirely.]] 1526 3. Consider the case of a screen reader used by blind Internet users 1527 who must listen to renderings of IDN domain names and possibly 1528 reproduce them on the keyboard. 1530 4. As a simplified example of this, assume one wanted to use a 1531 "heart" or "star" symbol in a label. This is problematic because 1532 those names are ambiguous in the Unicode system of naming (the 1533 actual Unicode names require far more qualification). A user or 1534 would-be registrant has no way to know -- absent careful study of 1535 the code tables -- whether it is ambiguous (e.g., where there are 1536 multiple "heart" characters) or not. Conversely, the user seeing 1537 the hypothetical label doesn't know whether to read it -- try to 1538 transmit it to a colleague by voice -- as "heart", as "love", as 1539 "black heart", or as any of the other examples below. 1541 5. The actual situation is even worse than this. There is no 1542 possible way for a normal, casual, user to tell the difference 1543 between the hearts of U+2665 and U+2765 and the stars of U+2606 1544 and U+2729 or the without somehow knowing to look for a 1545 distinction. We have a white heart (U+2661) and few black 1546 hearts. Consequently, describing a label as containing a heart 1547 hopelessly ambiguous: we can only know that it contains one of 1548 several characters that look like hearts or have "heart" in their 1549 names. In cities where "Square" is a popular part of a location 1550 name, one might well want to use a square symbol in a label as 1551 well and there are far more squares of various flavors in Unicode 1552 than there are hearts or stars. 1554 The consequence of these ambiguities is that symbols are a very poor 1555 basis for reliable communication. Consistent with this conclusion, 1556 the Unicode standard recommends that strings used in identifiers not 1557 contain symbols or punctuation [Unicode-UAX31]. Of course, these 1558 difficulties with symbols do not arise with actual pictographic 1559 languages and scripts which would be treated like any other language 1560 characters; the two should not be confused. 1562 7.7. Migration Between Unicode Versions: Unassigned Code Points 1564 In IDNA2003, labels containing unassigned code points are looked up 1565 on the assumption that, if they appear in labels and can be mapped 1566 and then resolved, the relevant standards must have changed and the 1567 registry has properly allocated only assigned values. 1569 In the protocol described in these documents, strings containing 1570 unassigned code points must not be either looked up or registered. 1571 In summary, the status of an unassigned character with regard to the 1572 DISALLOWED, PROTOCOL-VALID, and CONTEXTUAL RULE REQUIRED categories 1573 cannot be evaluated until a character is actually assigned and known. 1574 There are several reasons for this, with the most important ones 1575 being: 1577 o Tests involving the context of characters (e.g., some characters 1578 being permitted only adjacent to others of specific types) and 1579 integrity tests on complete labels are needed. Unassigned code 1580 points cannot be permitted because one cannot determine whether 1581 particular code points will require contextual rules (and what 1582 those rules should be) before characters are assigned to them and 1583 the properties of those characters fully understood. 1585 o It cannot be known in advance, and with sufficient reliability, 1586 that a no newly-assigned code point will associated with a 1587 character that would be disallowed by the rules in 1588 [IDNA2008-Tables] (such as a compatibility character). In 1589 IDNA2003, since there is no direct dependency on NFKC (many of the 1590 entries in Stringprep's tables are based on NFKC, but IDNA2003 1591 depends only on Stringprep), allocation of a compatibility 1592 character might produce some odd situations, but it would not be a 1593 problem. In IDNA2008, where compatibility characters are 1594 DISALLOWED unless character-specific exceptions are made, 1595 permitting strings containing unassigned characters to be looked 1596 up would violate the principle that characters in DISALLOWED are 1597 not looked up. 1599 o The Unicode Standard specifies that an unassigned code point 1600 normalizes (and, where relevant, case folds) to itself. If the 1601 code point is later assigned to a character, and particularly if 1602 the newly-assigned code point has a combining class that 1603 determines its placement relative to other combining characters, 1604 it could normalize to some other code point or sequence. 1606 It is possible to argue that the issues above are not important and 1607 that, as a consequence, it is better to retain the principle of 1608 looking up labels even if they contain unassigned characters because 1609 all of the important scripts and characters have been coded as of 1610 Unicode 5.1 and hence unassigned code points will be assigned only to 1611 obscure characters or archaic scripts. Unfortunately, that does not 1612 appear to be a safe assumption for at least two reasons. First, much 1613 the same claim of completeness has been made for earlier versions of 1614 Unicode. The reality is that a script that is obscure to much of the 1615 world may still be very important to those who use it. Cultural and 1616 linguistic preservation principles make it inappropriate to declare 1617 the script of no importance in IDNs. Second, we already have 1618 counterexamples in, e.g., the relationships associated with new Han 1619 characters being added (whether in the BMP or in Unicode Plane 2). 1621 Independent of the technical transition issues identified above, it 1622 can be observed that any addition of characters to an existing script 1623 to make it easier to use or to better accommodate particular 1624 languages may lead to transition issues. Such changes may change the 1625 preferred form for writing a particular string, changes that may be 1626 reflected, e.g., in keyboard transition modules that would 1627 necessarily be different from those for earlier versions of Unicode 1628 where the newer characters may not exist. This creates an inherent 1629 transition problem because attempts to access labels may use either 1630 the old or the new conventions, requiring registry action whether the 1631 older conventions were used in labels or not. The need to consider 1632 transition mechanisms is inherent to evolution of Unicode to better 1633 accommodate writing systems and is independent of how IDNs are 1634 represented in the DNS or how transitions among versions of those 1635 mechanisms occur. The requirement for transitions of this type is 1636 illustrated by the addition of Malayalam Chillu in Unicode 5.1.0. 1638 7.8. Other Compatibility Issues 1640 The 2003 IDNA model includes several odd artifacts of the context in 1641 which it was developed. Many, if not all, of these are potential 1642 avenues for exploits, especially if the registration process permits 1643 "source" names (names that have not been processed through IDNA and 1644 Nameprep) to be registered. As one example, since the character 1645 Eszett, used in German, is mapped by IDNA2003 into the sequence "ss" 1646 rather than being retained as itself or prohibited, a string 1647 containing that character but that is otherwise in ASCII is not 1648 really an IDN (in the U-label sense defined above) at all. After 1649 Nameprep maps the Eszett out, the result is an ASCII string and so 1650 does not get an xn-- prefix, but the string that can be displayed to 1651 a user appears to be an IDN. The newer version of the protocol 1652 eliminates this artifact. A character is either permitted as itself 1653 or it is prohibited; special cases that make sense only in a 1654 particular linguistic or cultural context can be dealt with as 1655 localization matters where appropriate. 1657 8. Name Server Considerations 1658 8.1. Processing Non-ASCII Strings 1660 Existing DNS servers do not know the IDNA rules for handling non- 1661 ASCII forms of IDNs, and therefore need to be shielded from them. 1662 All existing channels through which names can enter a DNS server 1663 database (for example, master files (as described in RFC 1034) and 1664 DNS update messages [RFC2136]) are IDN-unaware because they predate 1665 IDNA. Other sections of this document provide the needed shielding 1666 by ensuring that internationalized domain names entering DNS server 1667 databases through such channels have already been converted to their 1668 equivalent ASCII A-label forms. 1670 Because of the distinction made between the algorithms for 1671 Registration and Lookup in [IDNA2008-Protocol] (a domain name 1672 containing only ASCII codepoints can not be converted to an A-label), 1673 there can not be more than one A-label form for any given U-label. 1675 As specified in RFC 2181 [RFC2181], the DNS protocol explicitly 1676 allows domain labels to contain octets beyond the ASCII range 1677 (0000..007F), and this document does not change that. However, 1678 although the interpretation of octets 0080..00FF is well-defined in 1679 the DNS, many application protocols support only ASCII labels and 1680 there is no defined interpretation of these non-ASCII octets as 1681 characters and, in particular, no interpretation of case-independent 1682 matching for them (see, e.g., [RFC4343]). If labels containing these 1683 octets are returned to applications, unpredictable behavior could 1684 result. The A-label form, which cannot contain those characters, is 1685 the only standard representation for internationalized labels in the 1686 DNS protocol. 1688 8.2. DNSSEC Authentication of IDN Domain Names 1690 DNS Security (DNSSEC) [RFC2535] is a method for supplying 1691 cryptographic verification information along with DNS messages. 1692 Public Key Cryptography is used in conjunction with digital 1693 signatures to provide a means for a requester of domain information 1694 to authenticate the source of the data. This ensures that it can be 1695 traced back to a trusted source, either directly or via a chain of 1696 trust linking the source of the information to the top of the DNS 1697 hierarchy. 1699 IDNA specifies that all internationalized domain names served by DNS 1700 servers that cannot be represented directly in ASCII MUST use the 1701 A-label form. Conversion to A-labels MUST be performed prior to a 1702 zone being signed by the private key for that zone. Because of this 1703 ordering, it is important to recognize that DNSSEC authenticates a 1704 domain name containing A-labels or conventional LDH-labels, not 1705 U-labels. In the presence of DNSSEC, no form of a zone file or query 1706 response that contains a U-label may be signed or the signature 1707 validated. 1709 One consequence of this for sites deploying IDNA in the presence of 1710 DNSSEC is that any special purpose proxies or forwarders used to 1711 transform user input into IDNs must be earlier in the lookup flow 1712 than DNSSEC authenticating nameservers for DNSSEC to work. 1714 8.3. Root and other DNS Server Considerations 1716 IDNs in A-label form will generally be somewhat longer than current 1717 domain names, so the bandwidth needed by the root servers is likely 1718 to go up by a small amount. Also, queries and responses for IDNs 1719 will probably be somewhat longer than typical queries historically, 1720 so EDNS0 [RFC2671] support may be more important (otherwise, queries 1721 and responses may be forced to go to TCP instead of UDP). 1723 9. Internationalization Considerations 1725 DNS labels and fully-qualified domain names provide mnemonics that 1726 assist in identifying and referring to resources on the Internet. 1727 IDNs expand the range of those mnemonics to include those based on 1728 languages and character sets other than Western European and Roman- 1729 derived ones. But domain "names" are not, in general, words in any 1730 language. The recommendations of the IETF policy on character sets 1731 and languages, (BCP 18 [RFC2277]) are applicable to situations in 1732 which language identification is used to provide language-specific 1733 contexts. The DNS is, by contrast, global and international and 1734 ultimately has nothing to do with languages. Adding languages (or 1735 similar context) to IDNs generally, or to DNS matching in particular, 1736 would imply context dependent matching in DNS, which would be a very 1737 significant change to the DNS protocol itself. It would also imply 1738 that users would need to identify the language associated with a 1739 particular label in order to look that label up. That knowledge is 1740 generally not available because many labels are not words in any 1741 language and some may be words in more than one. 1743 10. IANA Considerations 1745 This section gives an overview of IANA registries required for IDNA. 1746 The actual definitions of, and specifications for, the first two, 1747 which must be newly-created for IDNA2008, appear in 1748 [IDNA2008-Tables]. This document describes the registries but does 1749 not specify any IANA actions. 1751 10.1. IDNA Character Registry 1753 The distinction among the major categories "UNASSIGNED", 1754 "DISALLOWED", "PROTOCOL-VALID", and "CONTEXTUAL RULE REQUIRED" is 1755 made by special categories and rules that are integral elements of 1756 [IDNA2008-Tables]. While not normative, an IANA registry of 1757 characters and scripts and their categories, updated for each new 1758 version of Unicode and the characters it contains, will be convenient 1759 for programming and validation purposes. The details of this 1760 registry are specified in [IDNA2008-Tables]. 1762 10.2. IDNA Context Registry 1764 IANA will create and maintain a list of approved contextual rules for 1765 characters that are defined in the IDNA Character Registry list as 1766 requiring a Contextual Rule (i.e., the types of rule described in 1767 Section 3.1.2). The details for those rules appear in 1768 [IDNA2008-Tables]. 1770 10.3. IANA Repository of IDN Practices of TLDs 1772 This registry, historically described as the "IANA Language Character 1773 Set Registry" or "IANA Script Registry" (both somewhat misleading 1774 terms) is maintained by IANA at the request of ICANN. It is used to 1775 provide a central documentation repository of the IDN policies used 1776 by top level domain (TLD) registries who volunteer to contribute to 1777 it and is used in conjunction with ICANN Guidelines for IDN use. 1779 It is not an IETF-managed registry and, while the protocol changes 1780 specified here may call for some revisions to the tables, these 1781 specifications have no direct effect on that registry and no IANA 1782 action is required as a result. 1784 11. Security Considerations 1786 11.1. General Security Issues with IDNA 1788 This document is purely explanatory and informational and 1789 consequently introduces no new security issues. It would, of course, 1790 be a poor idea for someone to try to implement from it; such an 1791 attempt would almost certainly lead to interoperability problems and 1792 might lead to security ones. A discussion of security issues with 1793 IDNA, including some relevant history, appears in [IDNA2008-Defs]. 1795 12. Acknowledgments 1797 The editor and contributors would like to express their thanks to 1798 those who contributed significant early (pre-WG) review comments, 1799 sometimes accompanied by text, especially Mark Davis, Paul Hoffman, 1800 Simon Josefsson, and Sam Weiler. In addition, some specific ideas 1801 were incorporated from suggestions, text, or comments about sections 1802 that were unclear supplied by Vint Cerf, Frank Ellerman, Michael 1803 Everson, Asmus Freytag, Erik van der Poel, Michel Suignard, and Ken 1804 Whistler, although, as usual, they bear little or no responsibility 1805 for the conclusions the editor and contributors reached after 1806 receiving their suggestions. Thanks are also due to Vint Cerf, Lisa 1807 Dusseault, Debbie Garside, and Jefsey Morfin for conversations that 1808 led to considerable improvements in the content of this document. 1810 A meeting was held on 30 January 2008 to attempt to reconcile 1811 differences in perspective and terminology about this set of 1812 specifications between the design team and members of the Unicode 1813 Technical Consortium. The discussions at and subsequent to that 1814 meeting were very helpful in focusing the issues and in refining the 1815 specifications. The active participants at that meeting were (in 1816 alphabetic order as usual) Harald Alvestrand, Vint Cerf, Tina Dam, 1817 Mark Davis, Lisa Dusseault, Patrik Faltstrom (by telephone), Cary 1818 Karp, John Klensin, Warren Kumari, Lisa Moore, Erik van der Poel, 1819 Michel Suignard, and Ken Whistler. We express our thanks to Google 1820 for support of that meeting and to the participants for their 1821 contributions. 1823 Useful comments and text on the WG versions of the draft were 1824 received from many participants in the IETF "IDNABIS" WG and a number 1825 of document changes resulted from mailing list discussions made by 1826 that group. Marcos Sanz provided specific analysis and suggestions 1827 that were exceptionally helpful in refining the text, as did Vint 1828 Cerf, Mark Davis, Martin Duerst, Andrew Sullivan, and Ken Whistler. 1829 Lisa Dusseault provided extensive editorial suggestions during the 1830 spring of 2009, most of which were incorporated. 1832 13. Contributors 1834 While the listed editor held the pen, the core of this document and 1835 the initial WG version represents the joint work and conclusions of 1836 an ad hoc design team consisting of the editor and, in alphabetic 1837 order, Harald Alvestrand, Tina Dam, Patrik Faltstrom, and Cary Karp. 1838 In addition, there were many specific contributions and helpful 1839 comments from those listed in the Acknowledgments section and others 1840 who have contributed to the development and use of the IDNA 1841 protocols. 1843 14. References 1845 14.1. Normative References 1847 [ASCII] American National Standards Institute (formerly United 1848 States of America Standards Institute), "USA Code for 1849 Information Interchange", ANSI X3.4-1968, 1968. 1851 ANSI X3.4-1968 has been replaced by newer versions with 1852 slight modifications, but the 1968 version remains 1853 definitive for the Internet. 1855 [IDNA2008-Bidi] 1856 Alvestrand, H. and C. Karp, "An updated IDNA criterion for 1857 right to left scripts", July 2008, . 1860 [IDNA2008-Defs] 1861 Klensin, J., "Internationalized Domain Names for 1862 Applications (IDNA): Definitions and Document Framework", 1863 November 2008, . 1866 [IDNA2008-Protocol] 1867 Klensin, J., "Internationalized Domain Names in 1868 Applications (IDNA): Protocol", November 2008, . 1871 [IDNA2008-Tables] 1872 Faltstrom, P., "The Unicode Code Points and IDNA", 1873 July 2008, . 1876 A version of this document is available in HTML format at 1877 http://stupid.domain.name/idnabis/ 1878 draft-ietf-idnabis-tables-02.html 1880 [RFC3490] Faltstrom, P., Hoffman, P., and A. Costello, 1881 "Internationalizing Domain Names in Applications (IDNA)", 1882 RFC 3490, March 2003. 1884 [RFC3492] Costello, A., "Punycode: A Bootstring encoding of Unicode 1885 for Internationalized Domain Names in Applications 1886 (IDNA)", RFC 3492, March 2003. 1888 [Unicode-UAX15] 1889 The Unicode Consortium, "Unicode Standard Annex #15: 1890 Unicode Normalization Forms", March 2008, 1891 . 1893 [Unicode51] 1894 The Unicode Consortium, "The Unicode Standard, Version 1895 5.1.0", 2008. 1897 defined by: The Unicode Standard, Version 5.0, Boston, MA, 1898 Addison-Wesley, 2007, ISBN 0-321-48091-0, as amended by 1899 Unicode 5.1.0 1900 (http://www.unicode.org/versions/Unicode5.1.0/). 1902 14.2. Informative References 1904 [BIG5] Institute for Information Industry of Taiwan, "Computer 1905 Chinese Glyph and Character Code Mapping Table, Technical 1906 Report C-26", 1984. 1908 There are several forms and variations and a closely- 1909 related standard, CNS 11643. See the discussion in 1910 Chapter 3 of Lunde, K., CJKV Information Processing, 1911 O'Reilly & Associates, 1999 1913 [GB18030] "Chinese National Standard GB 18030-2000: Information 1914 Technology -- Chinese ideograms coded character set for 1915 information interchange -- Extension for the basic set.", 1916 2000. 1918 [RFC0810] Feinler, E., Harrenstien, K., Su, Z., and V. White, "DoD 1919 Internet host table specification", RFC 810, March 1982. 1921 [RFC0952] Harrenstien, K., Stahl, M., and E. Feinler, "DoD Internet 1922 host table specification", RFC 952, October 1985. 1924 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 1925 STD 13, RFC 1034, November 1987. 1927 [RFC1035] Mockapetris, P., "Domain names - implementation and 1928 specification", STD 13, RFC 1035, November 1987. 1930 [RFC1123] Braden, R., "Requirements for Internet Hosts - Application 1931 and Support", STD 3, RFC 1123, October 1989. 1933 [RFC2136] Vixie, P., Thomson, S., Rekhter, Y., and J. Bound, 1934 "Dynamic Updates in the Domain Name System (DNS UPDATE)", 1935 RFC 2136, April 1997. 1937 [RFC2181] Elz, R. and R. Bush, "Clarifications to the DNS 1938 Specification", RFC 2181, July 1997. 1940 [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and 1941 Languages", BCP 18, RFC 2277, January 1998. 1943 [RFC2535] Eastlake, D., "Domain Name System Security Extensions", 1944 RFC 2535, March 1999. 1946 [RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", 1947 RFC 2671, August 1999. 1949 [RFC2673] Crawford, M., "Binary Labels in the Domain Name System", 1950 RFC 2673, August 1999. 1952 [RFC2782] Gulbrandsen, A., Vixie, P., and L. Esibov, "A DNS RR for 1953 specifying the location of services (DNS SRV)", RFC 2782, 1954 February 2000. 1956 [RFC3454] Hoffman, P. and M. Blanchet, "Preparation of 1957 Internationalized Strings ("stringprep")", RFC 3454, 1958 December 2002. 1960 [RFC3491] Hoffman, P. and M. Blanchet, "Nameprep: A Stringprep 1961 Profile for Internationalized Domain Names (IDN)", 1962 RFC 3491, March 2003. 1964 [RFC3743] Konishi, K., Huang, K., Qian, H., and Y. Ko, "Joint 1965 Engineering Team (JET) Guidelines for Internationalized 1966 Domain Names (IDN) Registration and Administration for 1967 Chinese, Japanese, and Korean", RFC 3743, April 2004. 1969 [RFC3987] Duerst, M. and M. Suignard, "Internationalized Resource 1970 Identifiers (IRIs)", RFC 3987, January 2005. 1972 [RFC4290] Klensin, J., "Suggested Practices for Registration of 1973 Internationalized Domain Names (IDN)", RFC 4290, 1974 December 2005. 1976 [RFC4343] Eastlake, D., "Domain Name System (DNS) Case Insensitivity 1977 Clarification", RFC 4343, January 2006. 1979 [RFC4690] Klensin, J., Faltstrom, P., Karp, C., and IAB, "Review and 1980 Recommendations for Internationalized Domain Names 1981 (IDNs)", RFC 4690, September 2006. 1983 [RFC4713] Lee, X., Mao, W., Chen, E., Hsu, N., and J. Klensin, 1984 "Registration and Administration Recommendations for 1985 Chinese Domain Names", RFC 4713, October 2006. 1987 [Unicode-Security] 1988 The Unicode Consortium, "Unicode Technical Standard #39: 1989 Unicode Security Mechanisms", August 2008, 1990 . 1992 [Unicode-UAX31] 1993 The Unicode Consortium, "Unicode Standard Annex #31: 1994 Unicode Identifier and Pattern Syntax", March 2008, 1995 . 1997 [Unicode-UTR36] 1998 The Unicode Consortium, "Unicode Technical Report #36: 1999 Unicode Security Considerations", July 2008, 2000 . 2002 Appendix A. Change Log 2004 [[ RFC Editor: Please remove this appendix. ]] 2006 A.1. Changes between Version -00 and Version -01 of 2007 draft-ietf-idnabis-rationale 2009 o Clarified the U-label definition to note that U-labels must 2010 contain at least one non-ASCII character. Also clarified the 2011 relationship among label types. 2013 o Rewrote the discussion of Labels in Registration (Section 7.1.2) 2014 and related text about IDNA-validity (in the "Defs" document as of 2015 -04 of this one) to narrow its focus and remove more general 2016 restrictions. Added a temporary note in line to explain the 2017 situation. 2019 o Changed the "IDNA uses Unicode" statement to focus on 2020 compatibility with IDNA2003 and avoid more general or 2021 controversial assertions. 2023 o Added a discussion of examples to Section 7.1 2025 o Made a number of other small editorial changes and corrections 2026 suggested by Mark Davis. 2028 o Added several more discussion anchors and notes and expanded or 2029 updated some existing ones. 2031 A.2. Version -02 2033 o Trimmed change log, removing information about pre-WG drafts. 2035 o Adjusted discussion of Contextual Rules to match the new location 2036 of the tables and some conceptual material. 2038 o Rewrote the material on preprocessing somewhat. 2040 o Moved the material about relationships with IDNA2003 to be part of 2041 a single section on transitions. 2043 o Removed several placeholders and made editorial changes in 2044 accordance with decisions made at IETF 72 in Dublin and not 2045 disputed on the mailing list. 2047 A.3. Version -03 2049 This special update to the Rationale document is intended to try to 2050 get the discussion of what is normative or not under control. While 2051 the IETF does not normally annotate individual sections of documents 2052 with whether they are normative or not, concerns that we don't know 2053 which is which, claims that some material is normative that would be 2054 problematic if so classified, etc., argue that we should at least be 2055 able to have a clear discussion on the subject. 2057 Two annotations have been applied to sections that might reasonably 2058 be considered normative. One annotation is based on the list of 2059 sections in Mark Davis's note of 29 September (http:// 2060 www.alvestrand.no/pipermail/idna-update/2008-September/002667.html). 2061 The other is based on an elaboration of John Klensin's response on 7 2062 October (http://www.alvestrand.no/pipermail/idna-update/2008-October/ 2063 002691.html). These should just be considered two suggestions to 2064 illuminate and, one hopes, advance the Working Group's discussions. 2066 Some additional editorial changes have been made, but they are 2067 basically trivial. In the editor's judgment, it is not possible to 2068 make significantly more progress with this document until the matter 2069 of document organization is settled. 2071 A.4. Version -04 2073 o Definitional and other normative material moved to new document 2074 (draft-ietf-idnabis-defs). Version -03 annotations removed. 2076 o Material on differences between IDNA2003 and IDNA2008 moved to an 2077 appendix in Protocol. 2079 o Material left over from the origins of this document as a 2080 preliminary proposal has been removed or rewritten. 2082 o Changes made to reflect consensus call results, including removing 2083 several placeholder notes for discussion. 2085 o Added more material, including discussion of historic scripts, to 2086 Section 3.2 on registration policies. 2088 o Added a new section (Section 7.2) to contain specific discussion 2089 of handling of characters that are interpreted differently in 2090 input to IDNA2003 and 2008. 2092 o Some material, including this section/appendix, rearranged. 2094 A.5. Version -05 2096 o Many small editorial changes, including changes to eliminate the 2097 last vestiges of what appeared to be 2119 language (upper-case 2098 MUST, SHOULD, or MAY) and small adjustments to terminology. 2100 A.6. Version -06 2102 o Removed Security Considerations material and pointed to Defs, 2103 where it now appears as of version 05. 2105 o Started changing uses of "IDNA2008" in running text to "in these 2106 specifications" or the equivalent. These documents are titled 2107 simply "IDNA"; once they are standardized, "the current version" 2108 may be a more appropriate reference than one containing a year. 2109 As discussed on the mailing list, we can and should discuss how to 2110 refer to these documents at an appropriate time (e.g., when we 2111 know when we will be finished) but, in the interim, it seems 2112 appropriate to simply start getting rid of the version-specific 2113 terminology where it can naturally be removed. 2115 o Additional discussion of mappings, etc., especially for case- 2116 sensitivity. 2118 o Clarified relationship to base DNS specifications. 2120 o Consolidated discussion of lookup of unassigned characters. 2122 o More editorial fine-tuning. 2124 A.7. Version -07 2126 o Revised terminology by adding terms: NR-LDH-label, Invalid-A-label 2127 (or False-A-label), R-LDH-label, valid IDNA-label in 2128 Section 1.3.3. 2130 o Moved the "name server considerations" material to this document 2131 from Protocol because it is non-normative and not part of the 2132 protocol itself. 2134 o To improve clarity, redid discussion of the reasons why looking up 2135 unassigned code points is prohibited. 2137 o Editorial and other non-substantive corrections to reflect earlier 2138 errors as well as new definitions and terminology. 2140 A.8. Version -08 2142 o Slight revision to "contextual" discussion (Section 3.1.2) and 2143 moving it to a separate subsection, rather than under "PVALID", 2144 for better parallelism with Tables. Also reflected Mark's 2145 comments about the limitations of the approach. 2147 o Added placeholder notes as reminders of where references to the 2148 other documents need Section numbers. More of these will be added 2149 as needed (feel free to identify relevant places), but the actual 2150 section numbers will not be inserted until the documents are 2151 completely stable, i.e., on their way to the RFC Editor. 2153 A.9. Version -09 2155 o Small editorial changes to clarify transition possibilities. 2157 o Small clarification to the description of DNS "exact match". 2159 o Added discussion of adding characters to an existing script to the 2160 discussion of unassigned code point transitions in Section 7.7. 2162 o Tightened up the discussion of non-ASCII string processing 2163 (Section 8.1) slightly. 2165 o Removed some placeholders and comments that have been around long 2166 enough to be considered acceptable or that no longer seem 2167 necessary for other reasons. 2169 A.10. Version -10 2171 o Extensive editorial improvements, mostly due to suggestions from 2172 Lisa Dusseault. 2174 o Changes required for the new "mapping" approach and document have, 2175 in general, not been incorporated despite several suggestions. 2176 The editor intends to wait until the mapping model is stable, or 2177 at least until -11 of this document, before trying to incorporate 2178 those suggestions. 2180 Author's Address 2182 John C Klensin 2183 1770 Massachusetts Ave, Ste 322 2184 Cambridge, MA 02140 2185 USA 2187 Phone: +1 617 245 1457 2188 Email: john+ietf@jck.com