idnits 2.17.00 (12 Aug 2021) /tmp/idnits14622/draft-ietf-idnabis-rationale-04.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 16. -- Found old boilerplate from RFC 3978, Section 5.5, updated by RFC 4748 on line 2010. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 2021. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 2028. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 2034. 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 522: '...IGNED category. Such code points MUST...' RFC 2119 keyword, line 532: '...cies, registries SHOULD develop and ap...' RFC 2119 keyword, line 659: '..., and as an A-label. Applications MAY...' RFC 2119 keyword, line 662: '...s. In general, they SHOULD allow, but...' RFC 2119 keyword, line 683: '...ly allows one charset, the labels MUST...' (6 more instances...) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust Copyright Line does not match the current year -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (November 2, 2008) is 4947 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ASCII' is defined on line 1752, but no explicit reference was found in the text == Unused Reference: 'Unicode-UAX15' is defined on line 1793, but no explicit reference was found in the text == Unused Reference: 'RFC1034' is defined on line 1829, but no explicit reference was found in the text == Unused Reference: 'RFC1035' is defined on line 1832, but no explicit reference was found in the text == Unused Reference: 'RFC2181' is defined on line 1838, but no explicit reference was found in the text == Unused Reference: 'RFC2673' is defined on line 1844, but no explicit reference was found in the text == Unused Reference: 'Unicode-UTR36' is defined on line 1889, 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 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 (~~), 10 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group J. Klensin 3 Internet-Draft November 2, 2008 4 Intended status: Informational 5 Expires: May 6, 2009 7 Internationalized Domain Names for Applications (IDNA): Background, 8 Explanation, and Rationale 9 draft-ietf-idnabis-rationale-04.txt 11 Status of this Memo 13 By submitting this Internet-Draft, each author represents that any 14 applicable patent or other IPR claims of which he or she is aware 15 have been or will be disclosed, and any of which he or she becomes 16 aware will be disclosed, in accordance with Section 6 of BCP 79. 18 Internet-Drafts are working documents of the Internet Engineering 19 Task Force (IETF), its areas, and its working groups. Note that 20 other groups may also distribute working documents as Internet- 21 Drafts. 23 Internet-Drafts are draft documents valid for a maximum of six months 24 and may be updated, replaced, or obsoleted by other documents at any 25 time. It is inappropriate to use Internet-Drafts as reference 26 material or to cite them other than as "work in progress." 28 The list of current Internet-Drafts can be accessed at 29 http://www.ietf.org/ietf/1id-abstracts.txt. 31 The list of Internet-Draft Shadow Directories can be accessed at 32 http://www.ietf.org/shadow.html. 34 This Internet-Draft will expire on May 6, 2009. 36 Abstract 38 Several years have passed since the original protocol for 39 Internationalized Domain Names (IDNs) was completed and deployed. 40 During that time, a number of issues have arisen, including the need 41 to update the system to deal with newer versions of Unicode. Some of 42 these issues require tuning of the existing protocols and the tables 43 on which they depend. This document provides an overview of a 44 revised system and provides explanatory material for its components. 46 Table of Contents 48 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 49 1.1. Context and Overview . . . . . . . . . . . . . . . . . . . 4 50 1.2. Discussion Forum . . . . . . . . . . . . . . . . . . . . . 4 51 1.3. Terminology . . . . . . . . . . . . . . . . . . . . . . . 4 52 1.3.1. Documents and Standards . . . . . . . . . . . . . . . 4 53 1.3.2. DNS "Name" Terminology . . . . . . . . . . . . . . . . 5 54 1.3.3. New Terminology and Restrictions . . . . . . . . . . . 5 55 1.4. Objectives . . . . . . . . . . . . . . . . . . . . . . . . 6 56 1.5. Applicability and Function of IDNA . . . . . . . . . . . . 6 57 1.6. Comprehensibility of IDNA Mechanisms and Processing . . . 7 58 2. Processing in IDNA2008 . . . . . . . . . . . . . . . . . . . . 9 59 3. Permitted Characters: An Inclusion List . . . . . . . . . . . 9 60 3.1. A Tiered Model of Permitted Characters and Labels . . . . 9 61 3.1.1. PROTOCOL-VALID . . . . . . . . . . . . . . . . . . . . 10 62 3.1.2. DISALLOWED . . . . . . . . . . . . . . . . . . . . . . 11 63 3.1.3. UNASSIGNED . . . . . . . . . . . . . . . . . . . . . . 12 64 3.2. Registration Policy . . . . . . . . . . . . . . . . . . . 12 65 3.3. Layered Restrictions: Tables, Context, Registration, 66 Applications . . . . . . . . . . . . . . . . . . . . . . . 13 67 4. Issues that Constrain Possible Solutions . . . . . . . . . . . 13 68 4.1. Display and Network Order . . . . . . . . . . . . . . . . 13 69 4.2. Entry and Display in Applications . . . . . . . . . . . . 15 70 4.3. Linguistic Expectations: Ligatures, Digraphs, and 71 Alternate Character Forms . . . . . . . . . . . . . . . . 16 72 4.4. Case Mapping and Related Issues . . . . . . . . . . . . . 18 73 4.5. Right to Left Text . . . . . . . . . . . . . . . . . . . . 19 74 5. IDNs and the Robustness Principle . . . . . . . . . . . . . . 19 75 6. Front-end and User Interface Processing . . . . . . . . . . . 20 76 7. Migration from IDNA2003 and Unicode Version Synchronization . 23 77 7.1. Design Criteria . . . . . . . . . . . . . . . . . . . . . 23 78 7.1.1. General IDNA Validity Criteria . . . . . . . . . . . . 23 79 7.1.2. Labels in Registration . . . . . . . . . . . . . . . . 25 80 7.1.3. Labels in Lookup . . . . . . . . . . . . . . . . . . . 26 81 7.2. Changes in Character Interpretations . . . . . . . . . . . 27 82 7.3. More Flexibility in User Agents . . . . . . . . . . . . . 28 83 7.4. The Question of Prefix Changes . . . . . . . . . . . . . . 30 84 7.4.1. Conditions Requiring a Prefix Change . . . . . . . . . 30 85 7.4.2. Conditions Not Requiring a Prefix Change . . . . . . . 31 86 7.4.3. Implications of Prefix Changes . . . . . . . . . . . . 31 87 7.5. Stringprep Changes and Compatibility . . . . . . . . . . . 31 88 7.6. The Symbol Question . . . . . . . . . . . . . . . . . . . 32 89 7.7. Migration Between Unicode Versions: Unassigned Code 90 Points . . . . . . . . . . . . . . . . . . . . . . . . . . 33 91 7.8. Other Compatibility Issues . . . . . . . . . . . . . . . . 34 92 8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 35 93 9. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 36 94 10. Internationalization Considerations . . . . . . . . . . . . . 36 95 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 36 96 11.1. IDNA Character Registry . . . . . . . . . . . . . . . . . 36 97 11.2. IDNA Context Registry . . . . . . . . . . . . . . . . . . 37 98 11.3. IANA Repository of IDN Practices of TLDs . . . . . . . . . 37 99 12. Security Considerations . . . . . . . . . . . . . . . . . . . 37 100 12.1. General Security Issues with IDNA . . . . . . . . . . . . 37 101 12.2. Security Differences from IDNA2003 . . . . . . . . . . . . 37 102 13. References . . . . . . . . . . . . . . . . . . . . . . . . . . 38 103 13.1. Normative References . . . . . . . . . . . . . . . . . . . 38 104 13.2. Informative References . . . . . . . . . . . . . . . . . . 39 105 Appendix A. Change Log . . . . . . . . . . . . . . . . . . . . . 41 106 A.1. Changes between Version -00 and Version -01 of 107 draft-ietf-idnabis-rationale . . . . . . . . . . . . . . . 41 108 A.2. Version -02 . . . . . . . . . . . . . . . . . . . . . . . 42 109 A.3. Version -03 . . . . . . . . . . . . . . . . . . . . . . . 42 110 A.4. Version -04 . . . . . . . . . . . . . . . . . . . . . . . 42 111 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 43 112 Intellectual Property and Copyright Statements . . . . . . . . . . 44 114 1. Introduction 116 1.1. Context and Overview 118 The original standards for Internationalized Domain Names (IDNs) were 119 completed and deployed starting in 2003. Those standards are known 120 as Internationalized Domain Names in Applications (IDNA), taken from 121 the name of the highest level standard within the group, RFC 3490 122 [RFC3490]. After those standards were deployed, a number of issues 123 arose that called for a new version of the IDNA protocol and the 124 associated tables, including a subset of those described in a recent 125 IAB report [RFC4690] and the need to update the system to deal with 126 newer versions of Unicode. This document further explains the issues 127 that have been encountered when they are important to understanding 128 of the revised protocols. It also provides an overview of the new 129 IDNA model and explanatory material for it. Additional explanatory 130 material for the specific components of the proposals appears with 131 the associated documents. 133 1.2. Discussion Forum 135 [[ RFC Editor: please remove this section. ]] 137 IDNA2008 is being discussed in the IETF "idnabis" Working Group and 138 on the mailing list idna-update@alvestrand.no 140 1.3. Terminology 142 Terminology that is critical for understanding this document and the 143 rest of the documents that make up IDNA2008, appears in 144 [IDNA2008-Defs]. That document also contains roadmap to the IDNA2008 145 document collection. No attempt should be made to understand this 146 document without the definitions and concepts that appear there. 148 1.3.1. Documents and Standards 150 This document uses the term "IDNA2003" to refer to the set of 151 standards that make up and support the version of IDNA published in 152 2003, i.e., those commonly known as the IDNA base specification 153 [RFC3490], Nameprep [RFC3491], Punycode [RFC3492], and Stringprep 154 [RFC3454]. In this document, those names are used to refer, 155 conceptually, to the individual documents, with the base IDNA 156 specification called just "IDNA". 158 The term "IDNA2008" is used to refer to a new version of IDNA as 159 described in this document and in the documents described in the 160 document listing of [IDNA2008-Defs]. IDNA2008 is not dependent on 161 any of the IDNA2003 specifications other than the one for Punycode 162 encoding. References to "these specifications" are to the entire 163 set. 165 1.3.2. DNS "Name" Terminology 167 These documents depart from historical DNS terminology and usage in 168 one important respect. Over the years, the community has talked very 169 casually about "names" in the DNS, beginning with calling it "the 170 domain name system". That terminology is fine in the very precise 171 sense that the identifiers of the DNS do provide names for objects 172 and addresses. But, in the context of IDNs, the term has introduced 173 some confusion, confusion that has increased further as people have 174 begun to speak of DNS labels in terms of the words or phrases of 175 various natural languages. 177 Historically, many, perhaps most, of the "names" in the DNS have been 178 mnemonics to identify some particular concept, object, or 179 organization. They are typically derived from, or rooted in, some 180 language because most people think in language-based ways. But, 181 because they are mnemonics, they need not obey the orthographic 182 conventions of any language: it is not a requirement that it be 183 possible for them to be "words". 185 This distinction is important because the reasonable goal of an IDN 186 effort is not to be able to write the great Klingon (or language of 187 one's choice) novel in DNS labels but to be able to form a usefully 188 broad range of mnemonics in ways that are as natural as possible in a 189 very broad range of scripts. 191 1.3.3. New Terminology and Restrictions 193 These documents [IDNA2008-Defs] introduce new terminology, and 194 precise definitions, for the terms "U-labels", "A-labels", labels 195 that are "IDNA-valid", and an "LDH-label" (differing from an LDH- 196 conformant label or fully-qualified domain name). The also introduce 197 a restriction, for IDNA-conformant applications and DNS zones in 198 which IDNA is used, on strings used as labels that contain "--" in 199 the third and fourth positions, essentially requiring that such 200 strings be IDNA-valid. This restriction on strings containing "--" 201 is required for three reasons: 203 o to prevent confusion with pre-IDNA coding forms; 205 o to permit future extensions that would require changing the 206 prefix, no matter how unlikely those might be (see Section 7.4); 207 and 209 o to reduce the opportunities for attacks via the encoding system. 211 1.4. Objectives 213 The intent of the IDNA revision effort, and hence of this document 214 and the associated ones, is to increase the usability and 215 effectiveness of internationalized domain names (IDNs) while 216 preserving or strengthening the integrity of references that use 217 them. The original "hostname" character definitions (see, e.g., 218 [RFC0810]) struck a balance between the creation of useful mnemonics 219 and the introduction of parsing problems or general confusion in the 220 contexts in which domain names are used. The objective of IDNA2008 221 is to preserve that balance while expanding the character repertoire 222 to include extended versions of Roman-derived scripts and scripts 223 that are not Roman in origin. No work of this sort is able to 224 completely eliminate sources of visual or textual confusion: such 225 confusion is possible even under the original host naming rules where 226 only ASCII characters were permitted. However, through the 227 application of different techniques at different points (see 228 Section 3.3), it should be possible to keep problems to an acceptable 229 minimum. One consequence of this general objective is that the 230 desire of some user or marketing community to use a particular string 231 --whether the reason is to try to write sentences of particular 232 languages in the DNS, to express a facsimile of the symbol for a 233 brand, or for some other purpose-- is not a primary goal within the 234 context of applications in the domain name space. 236 1.5. Applicability and Function of IDNA 238 The IDNA specification solves the problem of extending the repertoire 239 of characters that can be used in domain names to include a large 240 subset of the Unicode repertoire. 242 IDNA does not extend the service offered by DNS to the applications. 243 Instead, the applications (and, by implication, the users) continue 244 to see an exact-match lookup service. Either there is a single 245 exactly-matching name or there is no match. This model has served 246 the existing applications well, but it requires, with or without 247 internationalized domain names, that users know the exact spelling of 248 the domain names that are to be typed into applications such as web 249 browsers and mail user agents. The introduction of the larger 250 repertoire of characters potentially makes the set of misspellings 251 larger, especially given that in some cases the same appearance, for 252 example on a business card, might visually match several Unicode code 253 points or several sequences of code points. 255 The IDNA standard does not require any applications to conform to it, 256 nor does it retroactively change those applications. An application 257 can elect to use IDNA in order to support IDN while maintaining 258 interoperability with existing infrastructure. If an application 259 wants to use non-ASCII characters in domain names, IDNA is the only 260 currently-defined option. Adding IDNA support to an existing 261 application entails changes to the application only, and leaves room 262 for flexibility in front-end processing and more specifically in the 263 user interface (see Section 6). 265 A great deal of the discussion of IDN solutions has focused on 266 transition issues and how IDNs will work in a world where not all of 267 the components have been updated. Proposals that were not chosen by 268 the original IDN Working Group would have depended on updating of 269 user applications, DNS resolvers, and DNS servers in order for a user 270 to apply an internationalized domain name in any form or coding 271 acceptable under that method. While processing must be performed 272 prior to or after access to the DNS, IDNA requires no changes to the 273 DNS protocol or any DNS servers or the resolvers on user's computers. 275 IDNA allows the graceful introduction of IDNs not only by avoiding 276 upgrades to existing infrastructure (such as DNS servers and mail 277 transport agents), but also by allowing some rudimentary use of IDNs 278 in applications by using the ASCII-encoded representation of the 279 labels containing non-ASCII characters. While such names are user- 280 unfriendly to read and type, and hence not optimal for user input, 281 they can be used as a last resort to allow rudimentary IDN usage. 282 For example, they might be the best choice for display if it were 283 known that relevant fonts were not available on the user's computer. 284 In order to allow user-friendly input and output of the IDNs and 285 acceptance of some characters as equivalent to those to be processed 286 according to the protocol, the applications need to be modified to 287 conform to this specification. 289 IDNA uses the Unicode character repertoire, for continuity with the 290 original version of IDNA. 292 1.6. Comprehensibility of IDNA Mechanisms and Processing 294 One of the major goals of this work is to improve the general 295 understanding of how IDNA works and what characters are permitted and 296 what happens to them. Comprehensibility and predictability to users 297 and registrants are themselves important motivations and design goals 298 for this effort. The effort includes some new terminology and a 299 revised and extended model, both covered in this section, and some 300 more specific protocol, processing, and table modifications. Details 301 of the latter appear in other documents (see [IDNA2008-Defs]). 303 Several issues are inherent in the application of IDNs and, indeed, 304 almost any other system that tries to handle international characters 305 and concepts. They range from the apparently trivial --e.g., one 306 cannot display a character for which one does not have a font 307 available locally-- to the more complex and subtle. Many people have 308 observed that internationalization is just a tool to enable effective 309 localization while permitting some global uniformity. Issues of 310 display, of exactly how various strings and characters are entered, 311 and so on are inherently issues about localization and user interface 312 design. 314 A protocol such as IDNA can only assume that such operations as data 315 entry and reconciliation of differences in character forms are 316 possible. It may make some recommendations about how display might 317 work when characters and fonts are not available, but they can only 318 be general recommendations and, because display functions are rarely 319 controlled by the types of applications that would call upon IDNA, 320 will rarely be very effective. 322 However, shifting responsibility for character mapping and other 323 adjustments from the protocol (where it was located in IDNA2003) to 324 the user interface or processing before invoking IDNA raises issues 325 about both what that processing should do and about compatibility for 326 references prepared in an IDNA2003 context. Those issues are 327 discussed in Section 6. 329 Operations for converting between local character sets and normalized 330 Unicode are part of this general set of user interface issues. The 331 conversion is obviously not required at all in a Unicode-native 332 system that maintains all strings in Normalization Form C (NFC). It 333 may, however, involve some complexity in a system that is not 334 Unicode-native, especially if the elements of the local character set 335 do not map exactly and unambiguously into Unicode characters or do so 336 in a way that is not completely stable over time. Perhaps more 337 important, if a label being converted to a local character set 338 contains Unicode characters that have no correspondence in that 339 character set, the application may have to apply special, locally- 340 appropriate, methods to avoid or reduce loss of information. 342 Depending on the system involved, the major difficulty may not lie in 343 the mapping but in accurately identifying the incoming character set 344 and then applying the correct conversion routine. If a local 345 operating system uses one of the ISO 8859 character sets or an 346 extensive national or industrial system such as GB18030 [GB18030] or 347 BIG5 [BIG5], one must correctly identify the character set in use 348 before converting to Unicode even though those character coding 349 systems are substantially or completely Unicode-compatible (i.e., all 350 of the code points in them have an exact and unique mapping to 351 Unicode code points). It may be even more difficult when the 352 character coding system in local use is based on conceptually 353 different assumptions than those used by Unicode about, e.g., about 354 font encodings used for publications in some Indic scripts. Those 355 differences may not easily yield unambiguous conversions or 356 interpretations even if each coding system is internally consistent 357 and adequate to represent the local language and script. 359 2. Processing in IDNA2008 361 These specifications separate Domain Name Registration and Lookup in 362 the protocol specification. Doing so reflects current practice in 363 which per-registry restrictions and special processing are applied at 364 registration time but not during lookup. Even more important in the 365 longer term, it facilitates incremental addition of permitted 366 character groups to avoid freezing on one particular version of 367 Unicode. 369 The actual registration and lookup protocols for IDNA2008 are 370 specified in [IDNA2008-Protocol]. 372 3. Permitted Characters: An Inclusion List 374 This section provides an overview of the model used to establish the 375 algorithm and character lists of [IDNA2008-Tables] and describes the 376 names and applicability of the categories used there. Note that the 377 inclusion of a character in the first category group does not imply 378 that it can be used indiscriminately; some characters are associated 379 with contextual rules that must be applied as well. 381 The information given in this section is provided to make the rules, 382 tables, and protocol easier to understand. It is not normative. The 383 normative generating rules appear in [IDNA2008-Tables] and the rules 384 that actually determine what labels can be registered or looked up 385 are in [IDNA2008-Protocol]. 387 3.1. A Tiered Model of Permitted Characters and Labels 389 Moving to an inclusion model requires respecifying the list of 390 characters that are permitted in IDNs. In IDNA2003, the role and 391 utility of characters are independent of context and fixed forever 392 (or until the standard is replaced). Making completely context- 393 independent rules globally has proven impractical because some 394 characters, especially those that are called "Join_Controls" in 395 Unicode, are needed to make reasonable use of some scripts but have 396 no visible effect(s) in others. IDNA2003 prohibited those types of 397 characters entirely. But the restrictions were much too severe to 398 permit an adequate range of mnemonics for terminology based on some 399 languages. The requirement to support those characters but limit 400 their use to very specific contexts was reinforced by the observation 401 that handling of particular characters across the languages that use 402 a script, or the use of similar or identical-looking characters in 403 different scripts, is less well understood than many people believed 404 it was several years ago. 406 Independently of the characters chosen (see next subsection), the 407 theory is to divide the characters that appear in Unicode into three 408 categories: 410 3.1.1. PROTOCOL-VALID 412 Characters identified as "PROTOCOL-VALID" (often abbreviated 413 "PVALID") are, in general, permitted by IDNA for all uses in IDNs. 414 Their use may be restricted by rules about the context in which they 415 appear or by other rules that apply to the entire label in which they 416 are to be embedded. For example, any label that contains a character 417 in this category that has a "right-to-left" property must be used in 418 context with the "Bidi" rules (see [IDNA2008-Bidi]). 420 The term "PROTOCOL-VALID" is used to stress the fact that the 421 presence of a character in this category does not imply that a given 422 registry need accept registrations containing any of the characters 423 in the category. Registries are still expected to apply judgment 424 about labels they will accept and to maintain rules consistent with 425 those judgments (see [IDNA2008-Protocol] and Section 3.3). 427 Characters that are placed in the "PROTOCOL-VALID" category are never 428 removed from it unless the code points themselves are removed from 429 Unicode (such removal would be inconsistent with the Unicode 430 stability principles (see [Unicode51], Appendix F) and hence should 431 never occur). 433 3.1.1.1. Contextual Rules 435 Some characters may be unsuitable for general use in IDNs but 436 necessary for the plausible support of some scripts. The two most 437 commonly-cited examples are the zero-width joiner and non-joiner 438 characters (ZWJ, U+200D and ZWNJ, U+200C), but provisions for 439 unambiguous labels may require that other characters be restricted to 440 particular contexts. For example, the ASCII hyphen is not permitted 441 to start or end a label, whether that label contains non-ASCII 442 characters or not. 444 These characters must not appear in IDNs without additional 445 restrictions, typically because they have no visible consequences in 446 most scripts but affect format or presentation in a few others or 447 because they are combining characters that are safe for use only in 448 conjunction with particular characters or scripts. In order to 449 permit them to be used at all, they are specially identified as 450 "CONTEXTUAL RULE REQUIRED" and, when adequately understood, 451 associated with a rule. In addition, the rule will define whether it 452 is to be applied on lookup as well as registration. A distinction is 453 made between characters that indicate or prohibit joining (known as 454 "CONTEXT-JOINER" or "CONTEXTJ") and other characters requiring 455 contextual treatment ("CONTEXT-OTHER" or "CONTEXTO"). Only the 456 former are fully tested at lookup time. 458 3.1.1.2. Rules and Their Application 460 The actual rules may be present or absent. If present, they may have 461 values of "True" (character may be used in any position in any 462 label), "False" (character may not be used in any label), or may be a 463 set of procedural rules that specify the context in which the 464 character is permitted. 466 Examples of descriptions of typical rules, stated informally and in 467 English, include "Must follow a character from Script XYZ", "Must 468 occur only if the entire label is in Script ABC", "Must occur only if 469 the previous and subsequent characters have the DFG property". 471 Because it is easier to identify these characters than to know that 472 they are actually needed in IDNs or how to establish exactly the 473 right rules for each one, a rule may have a null value in a given 474 version of the tables. Characters associated with null rules are not 475 permitted to appear in putative labels for either registration or 476 lookup. Of course, a later version of the tables might contain a 477 non-null rule. 479 The description of the syntax of the rules, and the rules themselves, 480 appears in [IDNA2008-Tables]. 482 3.1.2. DISALLOWED 484 Some characters are sufficiently problematic for use in IDNs that 485 they should be excluded for both registration and lookup (i.e., IDNA- 486 conforming applications performing name lookup should verify that 487 these characters are absent; if they are present, the label strings 488 should be rejected rather than converted to A-labels and looked up. 490 Of course, this category would include code points that had been 491 removed entirely from Unicode should such removals ever occur. 493 Characters that are placed in the "DISALLOWED" category are expected 494 to never be removed from it or reclassified. If a character is 495 classified as "DISALLOWED" in error and the error is sufficiently 496 problematic, the only recourse would be either to introduce a new 497 code point into Unicode and classify it as "PROTOCOL-VALID" or for 498 the IETF to accept the considerable costs of an incompatible change 499 and replace the relevant RFC with one containing appropriate 500 exceptions. 502 There is provision for exception cases but, in general, characters 503 are placed into "DISALLOWED" if they fall into one or more of the 504 following groups: 506 o The character is a compatibility equivalent for another character. 507 In slightly more precise Unicode terms, application of 508 normalization method NFKC to the character yields some other 509 character. 511 o The character is an upper-case form or some other form that is 512 mapped to another character by Unicode casefolding. 514 o The character is a symbol or punctuation form or, more generally, 515 something that is not a letter, digit, or a mark that is used to 516 form a letter or digit. 518 3.1.3. UNASSIGNED 520 For convenience in processing and table-building, code points that do 521 not have assigned values in a given version of Unicode are treated as 522 belonging to a special UNASSIGNED category. Such code points MUST 523 NOT appear in labels to be registered or looked up. The category 524 differs from DISALLOWED in that code points are moved out of it by 525 the simple expedient of being assigned in a later version of Unicode 526 (at which point, they are classified into one of the other categories 527 as appropriate). 529 3.2. Registration Policy 531 While these recommendations cannot and should not define registry 532 policies, registries SHOULD develop and apply additional restrictions 533 to reduce confusion and other problems. For example, it is generally 534 believed that labels containing characters from more than one script 535 are a bad practice although there may be some important exceptions to 536 that principle. Some registries may choose to restrict registrations 537 to characters drawn from a very small number of scripts. For many 538 scripts, the use of variant techniques such as those as described in 539 RFC 3843 [RFC3743] and RFC 4290 [RFC4290], and illustrated for 540 Chinese by the tables described in RFC 4713 [RFC4713] may be helpful 541 in reducing problems that might be perceived by users. 543 In general, users will benefit if registries only permit characters 544 from scripts that are well-understood by the registry or its 545 advisers. If a registry decides to reduce opportunities for 546 confusion by constructing policies that disallow characters used in 547 historic writing systems or characters whose use is restricted to 548 specialized, highly technical contexts, some relevant information may 549 be found in Section 2.4 "Specific Character Adjustments", Table 4 550 "Candidate Characters for Exclusion from Identifiers" of 551 [Unicode-UAX31] and Section 3.1. "General Security Profile for 552 Identifiers" in [Unicode-Security]. 554 It is worth stressing that these principles of policy development and 555 application apply at all levels of the DNS, not only, e.g., TLD 556 registrations and that even a trivial, "anything permitted that is 557 valid under the protocol" policy is helpful in that it helps users 558 and application developers know what to expect. 560 3.3. Layered Restrictions: Tables, Context, Registration, Applications 562 The essence of the character rules in IDNA2008 is based on the 563 realization that there is no single magic bullet for any of the 564 issues associated with a multiscript DNS. Instead, the 565 specifications define a variety of approaches that, together, 566 constitute multiple lines of defense against ambiguity in identifiers 567 and loss of referential integrity. The actual character tables are 568 the first mechanism, protocol rules about how those characters are 569 applied or restricted in context are the second, and those two in 570 combination constitute the limits of what can be done by a protocol 571 alone. As discussed in the previous section (Section 3.2), 572 registries are expected to restrict what they permit to be 573 registered, devising and using rules that are designed to optimize 574 the balance between confusion and risk on the one hand and maximum 575 expressiveness in mnemonics on the other. 577 In addition, there is an important role for user agents in warning 578 against label forms that appear unreasonable given their knowledge of 579 local contexts and conventions. Of course, no approach based on 580 naming or identifiers alone can protect against all threats. 582 4. Issues that Constrain Possible Solutions 584 4.1. Display and Network Order 586 The correct treatment of domain names requires a clear distinction 587 between Network Order (the order in which the code points are sent in 588 protocols) and Display Order (the order in which the code points are 589 displayed on a screen or paper). The order of labels in a domain 590 name that contains characters that are normally written right to left 591 is discussed in [IDNA2008-Bidi]. In particular, there are questions 592 about the order in which labels are displayed if left to right and 593 right to left labels are adjacent to each other, especially if there 594 are also multiple consecutive appearances of one of the types. The 595 decision about the display order is ultimately under the control of 596 user agents --including web browsers, mail clients, and the like-- 597 which may be highly localized. Even when formats are specified by 598 protocols, the full composition of an Internationalized Resource 599 Identifier (IRI) [RFC3987] or Internationalized Email address 600 contains elements other than the domain name. For example, IRIs 601 contain protocol identifiers and field delimiter syntax such as 602 "http://" or "mailto:" while email addresses contain the "@" to 603 separate local parts from domain names. User agents are not required 604 to use those protocol-based forms directly but often do so. While 605 display, parsing, and processing within a label is specified by the 606 normative documents in the IDNA2008 collection, the relationship 607 between fully-qualified domain names and internationalized labels is 608 unchanged from the base DNS specifications. Comments in this 609 document about such full domain names are explanatory or examples of 610 what might be done and must not be considered normative. 612 Questions remain about protocol constraints implying that the overall 613 direction of these strings will always be left to right (or right to 614 left) for an IRI or email address, or if they even should conform to 615 such rules. These questions also have several possible answers. 616 Should a domain name abc.def, in which both labels are represented in 617 scripts that are written right to left, be displayed as fed.cba or 618 cba.fed? An IRI for clear text web access would, in network order, 619 begin with "http://" and the characters will appear as 620 "http://abc.def" -- but what does this suggest about the display 621 order? When entering a URI to many browsers, it may be possible to 622 provide only the domain name and leave the "http://" to be filled in 623 by default, assuming no tail (an approach that does not work for 624 protocols other than HTTP or whatever is chosen as the default). The 625 natural display order for the typed domain name on a right to left 626 system is fed.cba. Does this change if a protocol identifier, tail, 627 and the corresponding delimiters are specified? 629 While logic, precedent, and reality suggest that these are questions 630 for user interface design, not IETF protocol specifications, 631 experience in the 1980s and 1990s with mixing systems in which domain 632 name labels were read in network order (left to right) and those in 633 which those labels were read right to left would predict a great deal 634 of confusion, and heuristics that sometimes fail, if each 635 implementation of each application makes its own decisions on these 636 issues. 638 Any version of IDNA, including the current one, must be written in 639 terms of the network (transmission on the wire) order of characters 640 in labels and for the labels in complete (fully-qualified) domain 641 names and must be quite precise about those relationships. While 642 some strong suggestions about display order would be desirable to 643 reduce the chances for inconsistent transcription of domain names 644 from printed form, such suggestions are beyond the scope of these 645 specifications. 647 4.2. Entry and Display in Applications 649 Applications can accept domain names using any character set or sets 650 desired by the application developer, specified by the operating 651 system, or dictated by other constraints, and can display domain 652 names in any character set or character coding system. That is, the 653 IDNA protocol does not affect the interface between users and 654 applications. 656 An IDNA-aware application can accept and display internationalized 657 domain names in two formats: the internationalized character set(s) 658 supported by the application (i.e., an appropriate local 659 representation of a U-label), and as an A-label. Applications MAY 660 allow the display of A-labels, but are encouraged to not do so except 661 as an interface for special purposes, possibly for debugging, or to 662 cope with display limitations. In general, they SHOULD allow, but 663 not encourage, user input of that label form. A-labels are opaque 664 and ugly and malicious variations on them are not easily detected by 665 users. Where possible, they should thus only be exposed to users and 666 in contexts in which they are absolutely needed. Because IDN labels 667 can be rendered either as A-labels or U-labels, the application may 668 reasonably have an option for the user to select the preferred method 669 of display; if it does, rendering the U-label should normally be the 670 default. 672 Domain names are often stored and transported in many places. For 673 example, they are part of documents such as mail messages and web 674 pages. They are transported in many parts of many protocols, such as 675 both the control commands of SMTP and associated the message body 676 parts, and in the headers and the body content in HTTP. It is 677 important to remember that domain names appear both in domain name 678 slots and in the content that is passed over protocols. 680 In protocols and document formats that define how to handle 681 specification or negotiation of charsets, labels can be encoded in 682 any charset allowed by the protocol or document format. If a 683 protocol or document format only allows one charset, the labels MUST 684 be given in that charset. Of course, not all charsets can properly 685 represent all labels. If a U-label cannot be displayed in its 686 entirety, the only choice (without loss of information) may be to 687 display the A-label. 689 In any place where a protocol or document format allows transmission 690 of the characters in internationalized labels, labels SHOULD be 691 transmitted using whatever character encoding and escape mechanism 692 the protocol or document format uses at that place. This provision 693 is intended to prevent situations in which, e.g., UTF-8 domain names 694 appear embedded in text that is otherwise in some other character 695 coding. 697 All protocols that use domain name slots already have the capacity 698 for handling domain names in the ASCII charset. Thus, A-labels can 699 inherently be handled by those protocols. 701 4.3. Linguistic Expectations: Ligatures, Digraphs, and Alternate 702 Character Forms 704 [[anchor14: There is some internal redundancy and repetition in the 705 material in this section. Specific suggestions about to reduce or 706 eliminate redundant text for -05 would be appreciated.]] 708 Users often have expectations about character matching or equivalence 709 that are based on their own languages and the orthography of those 710 languages. These expectations may not be consistent with forms or 711 actions that can be naturally accommodated in a character coding 712 system, especially if multiple languages are written using the same 713 script but using different conventions. A Norwegian user might 714 expect a label with the ae-ligature to be treated as the same label 715 as one using the Swedish spelling with a-diaeresis even though 716 applying that mapping to English would be astonishing to users. A 717 user in German might expect a label with an o-umlaut and a label that 718 had "oe" substituted, but was otherwise the same, treated as 719 equivalent even though that substitution would be a clear error in 720 Swedish. A Chinese user might expect automatic matching of 721 Simplified and Traditional Chinese characters, but applying that 722 matching for Korean or Japanese text would create considerable 723 confusion. For that matter, an English user might expect "theater" 724 and "theatre" to match. 726 Related issues arise because there are a number of languages written 727 with alphabetic scripts in which single phonemes are written using 728 two characters, termed a "digraph", for example, the "ph" in 729 "pharmacy" and "telephone". (Note that characters paired in this 730 manner can also appear consecutively without forming a digraph, as in 731 "tophat".) Certain digraphs are normally indicated typographically 732 by setting the two characters closer together than they would be if 733 used consecutively to represent different phonemes. Some digraphs 734 are fully joined as ligatures (strictly designating setting totally 735 without intervening white space, although the term is sometimes 736 applied to close set pairs). An example of this may be seen when the 737 word "encyclopaedia" is set with a U+00E6 LATIN SMALL LIGATURE AE 738 (and some would not consider that word correctly spelled unless the 739 ligature form was used or the "a" was dropped entirely). When these 740 ligature and digraph forms have the same interpretation across all 741 languages that use a given script, application of Unicode 742 normalization generally resolves the differences and causes them to 743 match. When they have different interpretations, any requirements 744 for matching must utilize other methods, presumably at the registry 745 level, or users must be educated to understand that matching will not 746 occur. 748 Difficulties arise from the fact that a given ligature may be a 749 completely optional typographic convenience for representing a 750 digraph in one language (as in the above example with some spelling 751 conventions), while in another language it is a single character that 752 may not always be correctly representable by a two-letter sequence 753 (as in the above example with different spelling conventions). This 754 can be illustrated by many words in the Norwegian language, where the 755 "ae" ligature is the 27th letter of a 29-letter extended Latin 756 alphabet. It is equivalent to the 28th letter of the Swedish 757 alphabet (also containing 29 letters), U+00E4 LATIN SMALL LETTER A 758 WITH DIAERESIS, for which an "ae" cannot be substituted according to 759 current orthographic standards. 761 That character (U+00E4) is also part of the German alphabet where, 762 unlike in the Nordic languages, the two-character sequence "ae" is 763 usually treated as a fully acceptable alternate orthography for the 764 "umlauted a" character. The inverse is however not true, and those 765 two characters cannot necessarily be combined into an "umlauted a". 766 This also applies to another German character, the "umlauted o" 767 (U+00F6 LATIN SMALL LETTER O WITH DIAERESIS) which, for example, 768 cannot be used for writing the name of the author "Goethe". It is 769 also a letter in the Swedish alphabet where, like the "a with 770 diaeresis", it cannot be correctly represented as "oe" and in the 771 Norwegian alphabet, where it is represented, not as "o with 772 diaeresis", but as "slashed o", U+00F8. 774 Some of the ligatures that have explicit code points in Unicode were 775 given special handling in IDNA2003 and now pose additional problems 776 in transition. See Section 7.2. 778 Additional cases with alphabets written right to left are described 779 in Section 4.5. 781 Whether ligatures and digraphs are to be treated as a sequence of 782 characters or as a single standalone one constitute a problem that 783 cannot be resolved solely by operating on scripts. They are, 784 however, a key concern in the IDN context. Their satisfactory 785 resolution will require support in policies set by registries, which 786 therefore need to be particularly mindful not just of this specific 787 issue, but of all other related matters that cannot be dealt with on 788 an exclusively algorithmic and global basis. 790 Just as with the examples of different-looking characters that may be 791 assumed to be the same, it is in general impossible to deal with 792 these situations in a system such as IDNA -- or with Unicode 793 normalization generally -- since determining what to do requires 794 information about the language being used, context, or both. 795 Consequently, these specifications make no attempt to treat these 796 combined characters in any special way. However, their existence 797 provides a prime example of a situation in which a registry that is 798 aware of the language context in which labels are to be registered, 799 and where that language sometimes (or always) treats the two- 800 character sequences as equivalent to the combined form, should give 801 serious consideration to applying a "variant" model [RFC3743] 802 [RFC4290], or to prohibiting registration of one the forms entirely, 803 to reduce the opportunities for user confusion and fraud that would 804 result from the related strings being registered to different 805 parties. 807 4.4. Case Mapping and Related Issues 809 In the DNS, ASCII letters are stored with their case preserved. 810 Matching during the query process is case-independent, but none of 811 the information that might be represented by choices of case has been 812 lost. That model has been accidentally helpful because, as people 813 have created DNS labels by catenating words (or parts of words) to 814 form labels, case has often been used to distinguish among components 815 and make the labels more memorable. 817 The solution of keeping the characters separate but doing matching 818 independent of case is not feasible with IDNA or any IDNA-like model 819 because the matching would then have to be done on the server rather 820 than have characters mapped on the client. That situation was 821 recognized in IDNA2003 and nothing in IDNA2008 fundamentally changes 822 it or could do so. In IDNA2003, all characters are case-folded and 823 mapped. That results in upper-case characters being mapped to lower- 824 case ones and in some other transformations of alternate forms of 825 characters, especially those that do not have (or did not have) 826 upper-case forms. For example, Greek Final Form Sigma (U+03C2) is 827 mapped to the medial form (U+03C3) and Eszett (German Sharp S, 828 U+00DF) is mapped to "ss". Neither of these mappings is reversible 829 because the upper case of U+03C3 is the Upper Case Sigma (U+03A3) and 830 "ss" is an ASCII string. IDNA2008 permits, at the risk of some 831 incompatibility, slightly more flexibility in this area by avoid case 832 folding and treating these characters as themselves. Approaches to 833 handling the incompatibility are discussed in Section 7.2. Although 834 information is lost in IDNA2003's ToASCII operation so that, in some 835 sense, Final Sigma Eszett cannot be represented in an IDN at all, its 836 guarantee of mapping when those characters are used as input can be 837 interpreted as violating one of the conditions discussed in 838 Section 7.4.1 and hence requiring a prefix change. The consensus was 839 to not make a prefix change in spite of this issue. Of course, had a 840 prefix change been made (at the costs discussed in Section 7.4.3) 841 there would have been several options, including, if desired, 842 assignment of the character to the CONTEXTUAL RULE REQUIRED category 843 and requiring that it only be used in carefully-selected contexts. 845 4.5. Right to Left Text 847 In order to be sure that the directionality of right to left text is 848 unambiguous, IDNA2003 required that any label in which right to left 849 characters appear both starts and ends with them, not include any 850 characters with strong left to right properties (which excludes other 851 alphabetic characters but permits European digits), and rejects any 852 other string that contains a right to left character. This is one of 853 the few places where the IDNA algorithms (both in IDNA2003 and in 854 IDAN2008) are required to examine an entire label, not just 855 individual characters. The algorithmic model used in IDNA2003 856 rejects the label when the final character in a right to left string 857 requires a combining mark in order to be correctly represented. 859 That prohibition is not acceptable for writing systems for languages 860 written with consonantal alphabets to which diacritical vocalic 861 systems are applied, and for languages with orthographies derived 862 from them where the combining marks may have different functionality. 863 In both cases the combining marks can be essential components of the 864 orthography. Examples of this are Yiddish, written with an extended 865 Hebrew script, and Dhivehi (the official language of Maldives) which 866 is written in the Thaana script (which is, in turn, derived from the 867 Arabic script). IDNA2008 removes the restriction on final combining 868 characters with a new set of rules for right to left scripts and 869 their characters. Those new rules are specified in [IDNA2008-Bidi]. 871 5. IDNs and the Robustness Principle 873 The model of IDNs described in this document can be seen as a 874 particular instance of the "Robustness Principle" that has been so 875 important to other aspects of Internet protocol design. This 876 principle is often stated as "Be conservative about what you send and 877 liberal in what you accept" (See, e.g., Section 1.2.2 of the 878 applications-layer Host Requirements specification [RFC1123]). For 879 IDNs to work well, not only must the protocol be carefully designed 880 and implemented, but zone administrators (registries) must have and 881 require sensible policies about what is registered -- conservative 882 policies -- and implement and enforce them. 884 Conversely, lookup applications are expected to reject labels that 885 clearly violate global (protocol) rules (no one has ever seriously 886 claimed that being liberal in what is accepted requires being 887 stupid). However, once one gets past such global rules and deals 888 with anything sensitive to script or locale, it is necessary to 889 assume that garbage has not been placed into the DNS, i.e., one must 890 be liberal about what one is willing to look up in the DNS rather 891 than guessing about whether it should have been permitted to be 892 registered. 894 As mentioned elsewhere, if a string cannot be successfully found in 895 the DNS after the lookup processing described here, it makes no 896 difference whether it simply wasn't registered or was prohibited by 897 some rule at the registry. Applications should, however, be 898 sensitive to the fact that, because of the possibility of DNS 899 wildcards, the ability to successfully resolve a name does not 900 guarantee that it was actually registered. 902 If lookup applications, as a user interface (UI) or other local 903 matter, decide to warn about some strings that are valid under the 904 global rules but that they perceive as dangerous, that is their 905 prerogative and we can only hope that the market (and maybe 906 regulators) will reinforce the good choices and discourage the poor 907 ones. In this context, a lookup application that decides a string 908 that is valid under the protocol is dangerous and refuses to look it 909 up is in violation of the protocols; one that is willing to look 910 something up, but warns against it, is exercising a local choice. 912 6. Front-end and User Interface Processing 914 Domain names may be identified and processed in many contexts. They 915 may be typed in by users either by themselves or embedded in an 916 identifier structured for a particular protocol or class of protocols 917 such a email addresses, URIs, or IRIs. They may occur in running 918 text or be processed by one system after being provided in another. 919 Systems may wish to try to normalize URLs so as to determine (or 920 guess) whether a reference is valid or two references point to the 921 same object without actually looking the objects up and comparing 922 them (that is necessary, not just a choice, for URI types that are 923 not intended to be resolved). Some of these goals may be more easily 924 and reliably satisfied than others. While there are strong arguments 925 for any domain name that is placed "on the wire" -- transmitted 926 between systems -- to be in the minimum-ambiguity forms of A-labels, 927 U-labels, or LDH-labels, it is inevitable that programs that process 928 domain names will encounter variant forms. 930 One source of such forms will be labels created under IDNA2003 931 because that protocol allowed labels that were transformed before 932 they were turned from native-character into ACE ("xn--...") format by 933 mapping some characters into other. One consequence of the 934 transformations was that, when the ToUnicode and ToASCII operations 935 of IDNA2003 were applied, ToUnicode(ToASCII(original-label)) often 936 did not produce the original label. IDNA2008 explicitly defines 937 A-labels and U-labels as different forms of the same abstract label, 938 forms that are stable when conversions are performed between them, 939 without mappings. A different way of explaining this is that there 940 are, today, domain names in files on the Internet that use characters 941 that cannot be represented directly in, or recovered from, (A-label) 942 domain names but for which interpretations are provided by IDNA2003. 943 There are two major categories of such characters, those that are 944 removed by NFKC normalization and those upper-case characters that 945 are mapped to lower-case (there are also a few characters that are 946 given special-case mapping treatment in Stringprep including lower- 947 case characters that are case-folded into other lower-case characters 948 or strings). 950 Other issues in domain name identification and processing arise 951 because IDNA2003 specified that several other characters be treated 952 as equivalent to the ASCII period (dot, full stop) character used as 953 a label separator. If a string that might be a domain name appears 954 in an arbitrary context (such as running text), it is difficult, even 955 with only ASCII characters, to know whether an actual domain name (or 956 a protocol parameter like a URI) is present and where it starts and 957 ends. When using Unicode, this gets even more difficult if treatment 958 of certain special characters (like the dot that separates labels in 959 a domain name) depends on context (e.g., prior knowledge of whether 960 the string represents a domain name or not). That knowledge is not 961 available if the primary heuristic for identifying the presence of 962 domain names in strings depends on the presence of dots separating 963 groups of characters with no intervening spaces. 964 [[anchor16: Above text is a substitute for an earlier (pre -01) 965 version and is hoped to be more clear. Comments and improvements 966 welcome.]] 968 As discussed elsewhere in this document, the IDNA2008 model removes 969 all of these mappings and interpretations, including the equivalence 970 of different forms of dots, from the protocol, discouraging such 971 mappings and leaving them, when necessary, to local processing. This 972 should not be taken to imply that local processing is optional or can 973 be avoided entirely, even if doing so might have been desirable in a 974 world without IDNA2003 IDNs in files and archives. Instead, unless 975 the program context is such that it is known that any IDNs that 976 appear will be either U-labels or A-labels, or that other forms can 977 safely be rejected, some local processing of apparent domain name 978 strings will be required, both to maintain compatibility with 979 IDNA2003 and to prevent user astonishment. Such local processing, 980 while not specified in this document or the associated ones, will 981 generally take one of two forms: 983 o Generic Preprocessing. 984 When the context in which the program or system that processes 985 domain names operates is global, a reasonable balance must be 986 found that is sensitive to the broad range of local needs and 987 assumptions while, at the same time, not sacrificing the needs of 988 one language, script, or user population to those of another. 990 For this case, the best practice will usually be to apply NFKC and 991 case-mapping (or, perhaps better yet, Stringprep itself), plus 992 dot-mapping where appropriate, to the domain name string prior to 993 applying IDNA. That practice will not only yield a reasonable 994 compromise of user experience with protocol requirements but will 995 be almost completely compatible with the various forms permitted 996 by IDNA2003. 998 o Highly Localized Preprocessing. 999 Unlike the case above, there will be some situations in which 1000 software will be highly localized for a particular environment and 1001 carefully adapted to the expectations of users in that 1002 environment. The many discussions about using the Internet to 1003 preserve and support local cultures suggest that these cases may 1004 be more common in the future than they have been so far. 1006 In these cases, we should avoid trying to tell implementers what 1007 they should do, if only because they are quite likely (and for 1008 good reason) to ignore us. We would assume that they would map 1009 characters that the intuitions of their users would suggest be 1010 mapped and would hope that they would do that mapping as early as 1011 possible, storing A-label or U-label forms in files and 1012 transporting only those forms between systems. One can imagine 1013 switches about whether some sorts of mappings occur, warnings 1014 before applying them or, in a slightly more extreme version of the 1015 approach taken in Internet Explorer version 7 (IE7), systems that 1016 utterly refuse to handle "strange" characters at all if they 1017 appear in U-label form. None of those local decisions are a 1018 threat to interoperability as long as (i) only U-labels and 1019 A-labels are used in interchange with systems outside the local 1020 environment, (ii) no character that would be valid in a U-label as 1021 itself is mapped to something else, (iii) any local mappings are 1022 applied as a preprocessing step (or, for conversions from U-labels 1023 or A-labels to presentation forms, postprocessing), not as part of 1024 IDNA processing proper, and (iv) appropriate consideration is 1025 given to labels that might have entered the environment in 1026 conformance to IDNA2003. 1028 In either case, it is vital that user interface designs and, where 1029 the interfaces are not sufficient, users, be aware that the only 1030 forms of domain names that this protocol anticipates will resolve 1031 globally or compare equal when crude methods (i.e., those not 1032 conforming to the strict definition of label equivalence given in 1033 [IDNA2008-Defs]) are used are those in which all native-script labels 1034 are in U-label form. Forms that assume mapping will occur, 1035 especially forms that were not valid under IDNA2003, may or may not 1036 function in predictable ways across all implementations. 1038 7. Migration from IDNA2003 and Unicode Version Synchronization 1040 7.1. Design Criteria 1042 As mentioned above and in RFC 4690, two key goals of the IDNA2008 1043 design are to enable applications to be agnostic about whether they 1044 are being run in environments supporting any Unicode version from 3.2 1045 onward and to permit incrementally adding new characters, character 1046 groups, scripts, and other character collections as they are 1047 incorporated into Unicode, without disruption and, in the long term, 1048 without "heavy" processes such as those involving IETF consensus. 1049 (An IETF consensus process is required by the IDNA2008 specifications 1050 and is expected to be required and used until significant experience 1051 accumulates with IDNA operations and new versions of Unicode.) The 1052 mechanisms that support this are outlined above and elsewhere in the 1053 IDNA2008 document set, but this section reviews them in a context 1054 that may be more helpful to those who need to understand the approach 1055 and make plans for it. 1057 7.1.1. General IDNA Validity Criteria 1059 The general criteria for a putative label, and the collection of 1060 characters that make it up, to be considered IDNA-valid are (the 1061 actual rules are rigorously defined in the "Protocol" and "Tables" 1062 documents): 1064 o The characters are "letters", marks needed to form letters, 1065 numerals, or other code points used to write words in some 1066 language. Symbols, drawing characters, and various notational 1067 characters are permanently excluded -- some because they are 1068 actively dangerous in URI, IRI, or similar contexts and others 1069 because there is no evidence that they are important enough to 1070 Internet operations or internationalization to justify expansion 1071 of domain names beyond the general principle of "letters, digits, 1072 and hyphen" and the complexities that would come with it 1073 (additional discussion and rationale for the symbol decision 1074 appears in Section 7.6). 1076 o Other than in very exceptional cases, e.g., where they are needed 1077 to write substantially any word of a given language, punctuation 1078 characters are excluded as well. The fact that a word exists is 1079 not proof that it should be usable in a DNS label and DNS labels 1080 are not expected to be usable for multiple-word phrases (although 1081 they are certainly not prohibited if the conventions and 1082 orthography of a particular language cause that to be possible). 1083 Even for English, very common constructions -- contractions like 1084 "don't" or "it's", names that are written with apostrophes such as 1085 "O'Reilly", or characters for which apostrophes are common 1086 substitutes cannot be represented in DNS labels. Words in English 1087 whose usually-preferred spellings include diacritical marks cannot 1088 be represented under the original hostname rules, but most can be 1089 represented if treated as IDNs. 1091 o Characters that are unassigned (have no character assignment at 1092 all) in the version of Unicode being used by the registry or 1093 application are not permitted, even on lookup. There are at least 1094 two reasons for this. 1096 * Tests involving the context of characters (e.g., some 1097 characters being permitted only adjacent to ones of specific 1098 types but otherwise invisible or very problematic for other 1099 reasons) and integrity tests on complete labels are needed. 1100 Unassigned code points cannot be permitted because one cannot 1101 determine whether particular code points will require 1102 contextual rules (and what those rules should be) before 1103 characters are assigned to them and the properties of those 1104 characters fully understood. 1106 * Unicode specifies that an unassigned code point normalizes (and 1107 case folds) to itself. If the code point is later assigned to 1108 a character, and particularly if the newly-assigned code point 1109 has a combining class that determines its placement relative to 1110 other combining characters, it could normalize to some other 1111 code point or sequence, creating confusion and/or violating 1112 other rules listed here. 1114 o Any character that is mapped to another character by a current 1115 version of NFKC is prohibited as input to IDNA (for either 1116 registration or lookup). With a few exceptions, this principle 1117 excludes any character mapped to another by Nameprep [RFC3491]. 1119 Tables used to identify the characters that are IDNA-valid are 1120 expected to be driven by the principles above, principles that are 1121 specified exactly in [IDNA2008-Tables]). The rules given there are 1122 normative, rather than being just an interpretation of the tables. 1124 7.1.2. Labels in Registration 1126 Anyone entering a label into a DNS zone must properly validate that 1127 label -- i.e., be sure that the criteria for that label are met -- in 1128 order for applications to work as intended. This principle is not 1129 new. For example, since the DNS was first deployed, zone 1130 administrators have been expected to verify that names meet 1131 "hostname" [RFC0952] where necessary for the expected applications. 1132 Later addition of special service location formats [RFC2782] imposed 1133 new requirements on zone administrators for the use of labels that 1134 conform to the requirements of those formats. For zones that will 1135 contain IDNs, support for Unicode version-independence requires 1136 restrictions on all strings placed in the zone. In particular, for 1137 such zones: 1139 o Any label that appears to be an A-label, i.e., any label that 1140 starts in "xn--", MUST be IDNA-valid, i.e., they MUST be valid 1141 A-labels, as discussed in Section 2 above. 1143 o The Unicode tables (i.e., tables of code points, character 1144 classes, and properties) and IDNA tables (i.e., tables of 1145 contextual rules such as those that appear in the Tables 1146 document), MUST be consistent on the systems performing or 1147 validating labels to be registered. Note that this does not 1148 require that tables reflect the latest version of Unicode, only 1149 that all tables used on a given system are consistent with each 1150 other. 1152 Under this model, a registry (or entity communicating with a registry 1153 to accomplish name registrations) will need to update its tables -- 1154 both the Unicode-associated tables and the tables of permitted IDN 1155 characters -- to enable a new script or other set of new characters. 1156 It will not be affected by newer versions of Unicode, or newly- 1157 authorized characters, until and unless it wishes to make those 1158 registrations. The zone administrator is also responsible -- under 1159 the protocol and to registrants and users -- for both checking as 1160 required by the protocol and verification that whatever policies it 1161 develops are complied with, whether those policies are for minimizing 1162 risks due to confusable characters and sequences, for preserving 1163 language or script integrity, or for other purposes. Those checking 1164 and verification procedures are more extensive than those that are is 1165 expected of applications systems that look names up. 1167 Systems looking up or resolving DNS labels, especially IDN DNS 1168 labels, MUST be able to assume that applicable registration rules 1169 were followed for names entered into the DNS. 1171 7.1.3. Labels in Lookup 1173 Anyone looking up a label in a DNS zone is required to 1175 o Maintain a consistent set of tables, as discussed above. As with 1176 registration, the tables need not reflect the latest version of 1177 Unicode but they must be consistent. 1179 o Validate the characters in labels to be looked up only to the 1180 extent of determining that the U-label does not contain either 1181 code points prohibited by IDNA (categorized as "DISALLOWED") or 1182 code points that are unassigned in its version of Unicode. 1184 o Validate the label itself for conformance with a small number of 1185 whole-label rules, notably verifying that there are no leading 1186 combining marks, that the "bidi" conditions are met if right to 1187 left characters appear, that any required contextual rules are 1188 available and that, if such rules are associated with Joiner 1189 Controls, they are tested. 1191 o Avoid validating other contextual rules about characters, 1192 including mixed-script label prohibitions, although such rules may 1193 be used to influence presentation decisions in the user interface. 1195 By avoiding applying its own interpretation of which labels are valid 1196 as a means of rejecting lookup attempts, the lookup application 1197 becomes less sensitive to version incompatibilities with the 1198 particular zone registry associated with the domain name. 1200 An application or client that processes names according to this 1201 protocol and then resolves them in the DNS will be able to locate any 1202 name that is validly registered, as long as its version of the 1203 Unicode-associated tables is sufficiently up-to-date to interpret all 1204 of the characters in the label. Messages to users should distinguish 1205 between "label contains an unallocated code point" and other types of 1206 lookup failures. A failure on the basis of an old version of Unicode 1207 may lead the user to a desire to upgrade to a newer version, but will 1208 have no other ill effects (this is consistent with behavior in the 1209 transition to the DNS when some hosts could not yet handle some forms 1210 of names or record types). 1212 7.2. Changes in Character Interpretations 1214 [[anchor19: Note in Draft: This subsection is completely new in 1215 version -04 of this document. It could almost certainly use 1216 improvement. It also contains some material that is redundant with 1217 material in other sections. I have not tried to remove that material 1218 and will not do so until the WG concludes that this section is 1219 relatively stable, but would appreciate help in identifying what 1220 should be removed or how this might be enhanced to contain more of 1221 that other material. --JcK]] 1223 In those scripts that make case distinctions, there are a few 1224 characters for which an obvious and unique upper case character has 1225 not historically been available to match a lower case one or vice 1226 versa. For those characters, the mappings used in constructing the 1227 Stringprep tables for IDNA2003, performed using the Unicode CaseFold 1228 operation (See Section 5.8 of the Unicode Standard [Unicode51]), 1229 generate different characters or sets of characters. Those 1230 operations are not reversible and lose even more information than 1231 traditional upper case or lower case transformations, but are more 1232 useful than those transformations for comparison purposes. Two 1233 notable characters of this type are the German character Eszett 1234 (Sharp S, U+00DF) and the Greek Final Form Sigma (U+03C2). The 1235 former is case-folded to the ASCII string "ss", the latter to a 1236 medial (Lower Case) Sigma (U+03C3). 1238 The decision to eliminate mappings, including case folding, from the 1239 IDNA2008 protocol in order to make A-labels and U-labels idempotent 1240 made these characters problematic. If they were to be disallowed, 1241 important words and mnemonics could not be written in 1242 orthographically reasonable ways. If they were to be permitted as 1243 characters distinct from the forms produced by case folding, there 1244 would be no information loss and registries would have maximum 1245 flexibility, but labels using those characters that were looked up 1246 according to IDNA2003 rules would be transformed into A-labels using 1247 their case-mapped variations while lookup according to IDNA2008 rules 1248 would be based on different A-labels that represented the actual 1249 characters. 1251 With the understanding that there would be incompatibility either way 1252 but a judgment that the incompatibility was not significant enough to 1253 just a prefix change, the WG concluded that Eszett and Final Form 1254 Sigma should be treated as distinct and Protocol-Valid characters. 1256 The decision faces registries, especially registries maintaining 1257 zones for third parties, with a variation on what has become a 1258 familiar problem: how to introduce a new service in a way that does 1259 not create confusion or significantly weaken or invalidate existing 1260 identifiers. 1262 While it is beyond the scope of these documents to specify a 1263 preference for any of them, or to suggest that there are not other 1264 possibilities, there have traditionally been several approaches to 1265 problems of this type: 1267 o Do not permit use of the newly-available character at the registry 1268 level. This might cause lookup failures if a domain name were 1269 written with the expectation of the IDNA2003 mapping behavior, but 1270 would eliminate any possibility of false matches. 1272 o Hold a "sunrise" arrangement in which holders of the previously- 1273 mapped labels (labels containing "ss" in the Eszett case or ones 1274 containing Lower Case Sigma in the Final Sigma case) are given 1275 priority (and perhaps other benefits) for registering the 1276 corresponding string containing the newly-available characters. 1278 o Adopt some sort of "variant" approach in which registrants either 1279 obtained labels with both character forms or one of them was 1280 blocked from registration by anyone but the registrant of the 1281 other form. 1283 In principle, lookup applications could also compensate for the 1284 difference in interpretation by looking up the string according to 1285 the IDNA208 interpretation and then, if that failed, doing the lookup 1286 with the mapping, simulating the IDNA2003 interpretation. The risk 1287 of false positives is such that this is generally to be discouraged 1288 unless the application is able to engage in a "did you really mean" 1289 dialogue with the end user. 1291 7.3. More Flexibility in User Agents 1293 These specifications do not perform mappings between one character or 1294 code point and others for any reason. Instead, they prohibit the 1295 characters that would be mapped to others by normalization, case 1296 folding (with exceptions for lower case characters that have no upper 1297 case form, which are retained), or other rules. As examples, while 1298 mathematical characters based on Latin ones are accepted as input to 1299 IDNA2003, they are prohibited in IDNA2008. Similarly, double-width 1300 characters and other variations are prohibited as IDNA input. 1302 Since the rules in [IDNA2008-Tables] have the effect that only 1303 strings that are not transformed by NFKC are valid, if an application 1304 chooses to perform NFKC normalization before lookup, that operation 1305 is safe since this will never make the application unable to look up 1306 any valid string. However, as discussed above, the application 1307 cannot guarantee that any other application will perform that 1308 mapping, so it should be used only with caution and for informed 1309 users. 1311 In many cases these prohibitions should have no effect on what the 1312 user can type as input to the lookup process. It is perfectly 1313 reasonable for systems that support user interfaces to perform some 1314 character mapping that is appropriate to the local environment. This 1315 would normally be done prior to actual invocation of IDNA. At least 1316 conceptually, the mapping would be part of the Unicode conversions 1317 discussed above and in [IDNA2008-Protocol]. However, those changes 1318 will be local ones only -- local to environments in which users will 1319 clearly understand that the character forms are equivalent. For use 1320 in interchange among systems, it appears to be much more important 1321 that U-labels and A-labels can be mapped back and forth without loss 1322 of information. 1324 One specific, and very important, instance of this strategy arises 1325 with case-folding. In the ASCII-only DNS, names are looked up and 1326 matched in a case-independent way, but no actual case-folding occurs. 1327 Names can be placed in the DNS in either upper or lower case form (or 1328 any mixture of them) and that form is preserved, returned in queries, 1329 and so on. IDNA2003 simulated that behavior for non-ASCII strings by 1330 performing case-folding at registration time (resulting in only 1331 lower-case IDNs in the DNS) and when names were looked up. 1333 As suggested earlier in this section, it appears to be desirable to 1334 do as little character mapping as possible consistent with having 1335 Unicode work correctly (e.g., NFC mapping to resolve different 1336 codings for the same character is still necessary although the 1337 specifications require that it be performed prior to invoking the 1338 protocol) and to make the mapping between A-labels and U-labels 1339 idempotent. Case-mapping is not an exception to this principle. If 1340 only lower case characters can be registered in the DNS (i.e., be 1341 present in a U-label), then IDNA2008 should prohibit upper-case 1342 characters as input. Some other considerations reinforce this 1343 conclusion. For example, an essential element of the ASCII case- 1344 mapping functions is that uppercase(character) must be equal to 1345 uppercase(lowercase(character)). That requirement may not be 1346 satisfied with IDNs. For example, there are some characters in 1347 scripts that use case distinction that do not have counterparts in 1348 one case or the other. The relationship between upper case and lower 1349 case may even be language-dependent, with different languages (or 1350 even the same language in different areas) expecting different 1351 mappings. Of course, the expectations of users who are accustomed to 1352 a case-insensitive DNS environment will probably be well-served if 1353 user agents perform case folding prior to IDNA processing, but the 1354 IDNA procedures themselves should neither require such mapping nor 1355 expect them when they are not natural to the localized environment. 1357 7.4. The Question of Prefix Changes 1359 The conditions that would require a change in the IDNA ACE prefix 1360 ("xn--" for the version of IDNA specified in [RFC3490]) have been a 1361 great concern to the community. A prefix change would clearly be 1362 necessary if the algorithms were modified in a manner that would 1363 create serious ambiguities during subsequent transition in 1364 registrations. This section summarizes our conclusions about the 1365 conditions under which changes in prefix would be necessary and the 1366 implications of such a change. 1368 7.4.1. Conditions Requiring a Prefix Change 1370 An IDN prefix change is needed if a given string would be looked up 1371 or otherwise interpreted differently depending on the version of the 1372 protocol or tables being used. Consequently, work to update IDNs 1373 would require a prefix change if, and only if, one of the following 1374 four conditions were met: 1376 1. The conversion of an A-label to Unicode (i.e., a U-label) yields 1377 one string under IDNA2003 (RFC3490) and a different string under 1378 IDNA2008. 1380 2. An input string that is valid under IDNA2003 and also valid under 1381 IDNA2008 yields two different A-labels with the different 1382 versions of IDNA. This condition is believed to be essentially 1383 equivalent to the one above except for a very small number of 1384 edge cases which may not, pragmatically, justify a prefix change 1385 (See Section 7.2). 1387 Note, however, that if the input string is valid under one 1388 version and not valid under the other, this condition does not 1389 apply. See the first item in Section 7.4.2, below. 1391 3. A fundamental change is made to the semantics of the string that 1392 is inserted in the DNS, e.g., if a decision were made to try to 1393 include language or specific script information in that string, 1394 rather than having it be just a string of characters. 1396 4. A sufficiently large number of characters is added to Unicode so 1397 that the Punycode mechanism for block offsets no longer has 1398 enough capacity to reference the higher-numbered planes and 1399 blocks. This condition is unlikely even in the long term and 1400 certain not to arise in the next few years. 1402 7.4.2. Conditions Not Requiring a Prefix Change 1404 In particular, as a result of the principles described above, none of 1405 the following changes require a new prefix: 1407 1. Prohibition of some characters as input to IDNA. This may make 1408 names that are now registered inaccessible, but does not require 1409 a prefix change. 1411 2. Adjustments in IDNA tables or actions, including normalization 1412 definitions, that affect characters that were already invalid 1413 under IDNA2003. 1415 3. Changes in the style of the IDNA definition that does not alter 1416 the actions performed by IDNA. 1418 7.4.3. Implications of Prefix Changes 1420 While it might be possible to make a prefix change, the costs of such 1421 a change are considerable. Even if they wanted to do so, all 1422 registries could not convert all IDNA2003 ("xn--") registrations to a 1423 new form at the same time and synchronize that change with 1424 applications supporting lookup. Unless all existing registrations 1425 were simply to be declared invalid (and perhaps even then) systems 1426 that needed to support both labels with old prefixes and labels with 1427 new ones would first process a putative label under the IDNA2008 1428 rules and try to look it up and then, if it were not found, would 1429 process the label under IDNA2003 rules and look it up again. That 1430 process could significantly slow down all processing that involved 1431 IDNs in the DNS especially since, in principle, a fully-qualified 1432 name could contain a mixture of labels that were registered with the 1433 old and new prefixes, a situation that would make the use of DNS 1434 caching very difficult. In addition, looking up the same input 1435 string as two separate A-labels would create some potential for 1436 confusion and attacks, since they could, in principle, map to 1437 different targets and then resolve to different DNS label nodes. 1439 Consequently, a prefix change is to be avoided if at all possible, 1440 even if it means accepting some IDNA2003 decisions about character 1441 distinctions as irreversible and/or giving special treatment to edge 1442 cases. 1444 7.5. Stringprep Changes and Compatibility 1446 The Nameprep [RFC3491] specification, a key part of IDNA2003, is a 1447 profile of Stringprep [RFC3454]. While Nameprep is a Stringprep 1448 profile specific to IDNA, Stringprep is used by a number of other 1449 protocols. Concerns have been expressed about problems for non-DNS 1450 uses of Stringprep being caused by changes to the specification 1451 intended to improve the handling of IDNs, most notably as this might 1452 affect identification and authentication protocols. The proposed new 1453 inclusion tables [IDNA2008-Tables], the reduction in the number of 1454 characters permitted as input for registration or lookup (Section 3), 1455 and even the proposed changes in handling of right to left strings 1456 [IDNA2008-Bidi] either give interpretations to strings prohibited 1457 under IDNA2003 or prohibit strings that IDNA2003 permitted. The 1458 IDNA2008 protocol does not use either Nameprep or Stringprep at all, 1459 so there are no side-effect changes to other protocols. 1461 It is particularly important to keep IDNA processing separate from 1462 processing for various security protocols because some of the 1463 constraints that are necessary for smooth and comprehensible use of 1464 IDNs may be unwanted or undesirable in other contexts. For example, 1465 the criteria for good passwords or passphrases are very different 1466 from those for desirable IDNs: passwords should be hard to guess, 1467 while domain names should normally be easily memorable. Similarly, 1468 internationalized SCSI identifiers and other protocol components are 1469 likely to have different requirements than IDNs. 1471 7.6. The Symbol Question 1473 One of the major differences between this specification and the 1474 original version of IDNA is that the original version permitted non- 1475 letter symbols of various sorts, including punctuation and line- 1476 drawing symbols, in the protocol. They were always discouraged in 1477 practice. In particular, both the "IESG Statement" about IDNA and 1478 all versions of the ICANN Guidelines specify that only language 1479 characters be used in labels. This specification disallows symbols 1480 entirely. There are several reasons for this, which include: 1482 o As discussed elsewhere, the original IDNA specification assumed 1483 that as many Unicode characters as possible should be permitted, 1484 directly or via mapping to other characters, in IDNs. This 1485 specification operates on an inclusion model, extrapolating from 1486 the LDH rules -- which have served the Internet very well -- to a 1487 Unicode base rather than an ASCII base. 1489 o Most Unicode names for letters are, in most cases, fairly 1490 intuitive, unambiguous and recognizable to users of the relevant 1491 script. Symbol names are more problematic because there may be no 1492 general agreement on whether a particular glyph matches a symbol; 1493 there are no uniform conventions for naming; variations such as 1494 outline, solid, and shaded forms may or may not exist; and so on. 1495 As just one example, consider a "heart" symbol as it might appear 1496 in a logo that might be read as "I love...". While the user might 1497 read such a logo as "I love..." or "I heart...", considerable 1498 knowledge of the coding distinctions made in Unicode is needed to 1499 know that there more than one "heart" character (e.g., U+2665, 1500 U+2661, and U+2765) and how to describe it. These issues are of 1501 particular importance if strings are expected to be understood or 1502 transcribed by the listener after being read out loud. 1503 [[anchor20: The above paragraph remains controversial as to 1504 whether it is valid. The WG will need to make a decision if this 1505 section is not dropped entirely.]] 1507 o As a simplified example of this, assume one wanted to use a 1508 "heart" or "star" symbol in a label. This is problematic because 1509 those names are ambiguous in the Unicode system of naming (the 1510 actual Unicode names require far more qualification). A user or 1511 would-be registrant has no way to know -- absent careful study of 1512 the code tables -- whether it is ambiguous (e.g., where there are 1513 multiple "heart" characters) or not. Conversely, the user seeing 1514 the hypothetical label doesn't know whether to read it -- try to 1515 transmit it to a colleague by voice -- as "heart", as "love", as 1516 "black heart", or as any of the other examples below. 1518 o The actual situation is even worse than this. There is no 1519 possible way for a normal, casual, user to tell the difference 1520 between the hearts of U+2665 and U+2765 and the stars of U+2606 1521 and U+2729 or the without somehow knowing to look for a 1522 distinction. We have a white heart (U+2661) and few black hearts. 1523 Consequently, describing a label as containing a heart hopelessly 1524 ambiguous: we can only know that it contains one of several 1525 characters that look like hearts or have "heart" in their names. 1526 In cities where "Square" is a popular part of a location name, one 1527 might well want to use a square symbol in a label as well and 1528 there are far more squares of various flavors in Unicode than 1529 there are hearts or stars. 1531 o The consequence of these ambiguities of description and 1532 dependencies on distinctions that were, or were not, made in 1533 Unicode codings is that symbols are a very poor basis for reliable 1534 communication. Consistent with this conclusion, the Unicode 1535 standard recommends that strings used in identifiers not contain 1536 symbols or punctuation [Unicode-UAX31]. Of course, these 1537 difficulties with symbols do not arise with actual pictographic 1538 languages and scripts which would be treated like any other 1539 language characters; the two should not be confused. 1541 7.7. Migration Between Unicode Versions: Unassigned Code Points 1543 In IDNA2003, labels containing unassigned code points are looked up 1544 on the theory that, if they appear in labels and can be mapped and 1545 then resolved, the relevant standards must have changed and the 1546 registry has properly allocated only assigned values. 1548 In IDNA2008, strings containing unassigned code points MUST NOT be 1549 either looked up or registered. There are several reasons for this, 1550 with the most important ones being: 1552 o It cannot be known with sufficient reliability in advance that a 1553 code point that was not previously assigned will not be assigned 1554 to a compatibility character. In IDNA2003, since there is no 1555 direct dependency on NFKC (Stringprep's tables are based on NFKC, 1556 but IDNA2003 depends only on Stringprep), allocation of a 1557 compatibility character might produce some odd situations, but it 1558 would not be a problem. In IDNA2008, where compatibility 1559 characters are generally assigned to DISALLOWED, permitting 1560 strings containing unassigned characters to be looked up would 1561 permit violating the principle that characters in DISALLOWED are 1562 not looked up. 1564 o More generally, the status of an unassigned character with regard 1565 to the DISALLOWED and PROTOCOL-VALID categories, and whether 1566 contextual rules are required with the latter, cannot be evaluated 1567 until a character is actually assigned and known. 1569 It is possible to argue that the issues above are not important and 1570 that, as a consequence, it is better to retain the principle of 1571 looking up labels even if they contain unassigned characters because 1572 all of the important scripts and characters have been coded as of 1573 Unicode 5.1 and hence unassigned code points will be assigned only to 1574 obscure characters or archaic scripts. Unfortunately, that does not 1575 appear to be a safe assumption for at least two reasons. First, much 1576 the same claim of completeness has been made for earlier versions of 1577 Unicode. The reality is that a script that is obscure to much of the 1578 world may still be very important to those who use it. Cultural and 1579 linguistic preservation principles make it inappropriate to declare 1580 the script of no importance in IDNs. Second, we already have 1581 counterexamples in, e.g., the relationships associated with new Han 1582 characters being added (whether in the BMP or in Unicode Plane 2). 1584 7.8. Other Compatibility Issues 1586 The existing (2003) IDNA model includes several odd artifacts of the 1587 context in which it was developed. Many, if not all, of these are 1588 potential avenues for exploits, especially if the registration 1589 process permits "source" names (names that have not been processed 1590 through IDNA and Nameprep) to be registered. As one example, since 1591 the character Eszett, used in German, is mapped by IDNA2003 into the 1592 sequence "ss" rather than being retained as itself or prohibited, a 1593 string containing that character but that is otherwise in ASCII is 1594 not really an IDN (in the U-label sense defined above) at all. After 1595 Nameprep maps the Eszett out, the result is an ASCII string and so 1596 does not get an xn-- prefix, but the string that can be displayed to 1597 a user appears to be an IDN. The proposed IDNA2008 eliminates this 1598 artifact. A character is either permitted as itself or it is 1599 prohibited; special cases that make sense only in a particular 1600 linguistic or cultural context can be dealt with as localization 1601 matters where appropriate. 1603 8. Acknowledgments 1605 The editor and contributors would like to express their thanks to 1606 those who contributed significant early (pre-WG) review comments, 1607 sometimes accompanied by text, especially Mark Davis, Paul Hoffman, 1608 Simon Josefsson, and Sam Weiler. In addition, some specific ideas 1609 were incorporated from suggestions, text, or comments about sections 1610 that were unclear supplied by Frank Ellerman, Michael Everson, Asmus 1611 Freytag, Erik van der Poel, Michel Suignard, and Ken Whistler, 1612 although, as usual, they bear little or no responsibility for the 1613 conclusions the editor and contributors reached after receiving their 1614 suggestions. Thanks are also due to Vint Cerf, Debbie Garside, and 1615 Jefsey Morphin for conversations that led to considerable 1616 improvements in the content of this document. 1618 A meeting was held on 30 January 2008 to attempt to reconcile 1619 differences in perspective and terminology about this set of 1620 specifications between the design team and members of the Unicode 1621 Technical Consortium. The discussions at and subsequent to that 1622 meeting were very helpful in focusing the issues and in refining the 1623 specifications. The active participants at that meeting were (in 1624 alphabetic order as usual) Harald Alvestrand, Vint Cerf, Tina Dam, 1625 Mark Davis, Lisa Dusseault, Patrik Faltstrom (by telephone), Cary 1626 Karp, John Klensin, Warren Kumari, Lisa Moore, Erik van der Poel, 1627 Michel Suignard, and Ken Whistler. We express our thanks to Google 1628 for support of that meeting and to the participants for their 1629 contributions. 1631 Useful comments and text on the WG versions of the draft were 1632 received from many participants in the IETF "IDNABIS" WG and a number 1633 of document changes resulted from mailing list discussions made by 1634 that group. Marcos Sanz provided specific analysis and suggestions 1635 that were exceptionally helpful in refining the text, as did Mark 1636 Davis, Martin Duerst, Ken Whistler, and Andrew Sullivan. 1638 9. Contributors 1640 While the listed editor held the pen, this core of this document and 1641 the initial WG version represents the joint work and conclusions of 1642 an ad hoc design team consisting of the editor and, in alphabetic 1643 order, Harald Alvestrand, Tina Dam, Patrik Faltstrom, and Cary Karp. 1644 In addition, there were many specific contributions and helpful 1645 comments from those listed in the Acknowledgments section and others 1646 who have contributed to the development and use of the IDNA 1647 protocols. 1649 10. Internationalization Considerations 1651 DNS labels and fully-qualified domain names provide mnemonics that 1652 assist in identifying and referring to resources on the Internet. 1653 IDNs expand the range of those mnemonics to include those based on 1654 languages and character sets other than Western European and Roman- 1655 derived ones. But domain "names" are not, in general, words in any 1656 language. The recommendations of the IETF policy on character sets 1657 and languages, BCP 18 [RFC2277] are applicable to situations in which 1658 language identification is used to provide language-specific 1659 contexts. The DNS is, by contrast, global and international and 1660 ultimately has nothing to do with languages. Adding languages (or 1661 similar context) to IDNs generally, or to DNS matching in particular, 1662 would imply context dependent matching in DNS, which would be a very 1663 significant change to the DNS protocol itself. It would also imply 1664 that users would need to identify the language associated with a 1665 particular label in order to look that label up, a decision that 1666 would be impossible in many or most cases. 1668 11. IANA Considerations 1670 This section gives an overview of registries required for IDNA. The 1671 actual definitions of the first two appear in [IDNA2008-Tables]. 1673 11.1. IDNA Character Registry 1675 The distinction among the three major categories "UNASSIGNED", 1676 "DISALLOWED", and "PROTOCOL-VALID" is made by special categories and 1677 rules that are integral elements of [IDNA2008-Tables]. Convenience 1678 in programming and validation requires a registry of characters and 1679 scripts and their categories, updated for each new version of Unicode 1680 and the characters it contains. The details of this registry are 1681 specified in [IDNA2008-Tables]. 1683 11.2. IDNA Context Registry 1685 For characters that are defined in the IDNA Character Registry list 1686 as PROTOCOL-VALID but requiring a contextual rule (i.e., the types of 1687 rule described in Section 3.1.1.1), IANA will create and maintain a 1688 list of approved contextual rules. The details for those rules 1689 appear in [IDNA2008-Tables]. 1691 11.3. IANA Repository of IDN Practices of TLDs 1693 This registry, historically described as the "IANA Language Character 1694 Set Registry" or "IANA Script Registry" (both somewhat misleading 1695 terms) is maintained by IANA at the request of ICANN. It is used to 1696 provide a central documentation repository of the IDN policies used 1697 by top level domain (TLD) registries who volunteer to contribute to 1698 it and is used in conjunction with ICANN Guidelines for IDN use. 1700 It is not an IETF-managed registry and, while the protocol changes 1701 specified here may call for some revisions to the tables, these 1702 specifications have no direct effect on that registry and no IANA 1703 action is required as a result. 1705 12. Security Considerations 1707 12.1. General Security Issues with IDNA 1709 This document in the IDNA2008 series is purely explanatory and 1710 informational and consequently introduces no new security issues. It 1711 would, of course, be a poor idea for someone to try to implement from 1712 it; such an attempt would almost certainly lead to interoperability 1713 problems and might lead to security ones. A discussion of security 1714 issues with IDNA2008, and IDNA generally, appears in [IDNA2008-Defs]. 1716 12.2. Security Differences from IDNA2003 1718 The registration and lookup models described in this set of documents 1719 change the mechanisms available for lookup applications to determine 1720 the validity of labels they encounter. In some respects, the ability 1721 to test is strengthened. For example, putative labels that contain 1722 unassigned code points will now be rejected, while IDNA2003 permitted 1723 them (something that is now recognized as a considerable source of 1724 risk). On the other hand, the protocol specification no longer 1725 assumes that the application that looks up a name will be able to 1726 determine, and apply, information about the protocol version used in 1727 registration. In theory, that may increase risk since the 1728 application will be able to do less pre-lookup validation. In 1729 practice, the protection afforded by that test has been largely 1730 illusory for reasons explained in RFC 4690 and above. 1732 Any change to Stringprep or, more broadly, the IETF's model of the 1733 use of internationalized character strings in different protocols, 1734 creates some risk of inadvertent changes to those protocols, 1735 invalidating deployed applications or databases, and so on. The same 1736 considerations that would require changing the IDN prefix (see the 1737 discussion of prefix changes in Section 7.4) are the ones that would, 1738 e.g., invalidate certificates or hashes that depend on Stringprep, 1739 but those cases require careful consideration and evaluation. More 1740 important, it is not necessary to change Stringprep at all in order 1741 to create a definition or implementation of IDNA as specified in this 1742 set of documents. Because these documents do not depend on 1743 Stringprep at all, the question of upgrading other protocols that do 1744 depend on Stringprep can be left to experts on those protocols: there 1745 is no dependency between IDNA changes and possible upgrades to 1746 security protocols or conventions. 1748 13. References 1750 13.1. Normative References 1752 [ASCII] American National Standards Institute (formerly United 1753 States of America Standards Institute), "USA Code for 1754 Information Interchange", ANSI X3.4-1968, 1968. 1756 ANSI X3.4-1968 has been replaced by newer versions with 1757 slight modifications, but the 1968 version remains 1758 definitive for the Internet. 1760 [IDNA2008-Bidi] 1761 Alvestrand, H. and C. Karp, "An updated IDNA criterion for 1762 right to left scripts", July 2008, . 1765 [IDNA2008-Defs] 1766 Klensin, J., "Internationalized Domain Names for 1767 Applications (IDNA): Definitions and Document Framework", 1768 November 2008, . 1771 [IDNA2008-Protocol] 1772 Klensin, J., "Internationalized Domain Names in 1773 Applications (IDNA): Protocol", November 2008, . 1776 [IDNA2008-Tables] 1777 Faltstrom, P., "The Unicode Code Points and IDNA", 1778 July 2008, . 1781 A version of this document is available in HTML format at 1782 http://stupid.domain.name/idnabis/ 1783 draft-ietf-idnabis-tables-02.html 1785 [RFC3490] Faltstrom, P., Hoffman, P., and A. Costello, 1786 "Internationalizing Domain Names in Applications (IDNA)", 1787 RFC 3490, March 2003. 1789 [RFC3492] Costello, A., "Punycode: A Bootstring encoding of Unicode 1790 for Internationalized Domain Names in Applications 1791 (IDNA)", RFC 3492, March 2003. 1793 [Unicode-UAX15] 1794 The Unicode Consortium, "Unicode Standard Annex #15: 1795 Unicode Normalization Forms", March 2008, 1796 . 1798 [Unicode51] 1799 The Unicode Consortium, "The Unicode Standard, Version 1800 5.1.0", 2008. 1802 defined by: The Unicode Standard, Version 5.0, Boston, MA, 1803 Addison-Wesley, 2007, ISBN 0-321-48091-0, as amended by 1804 Unicode 5.1.0 1805 (http://www.unicode.org/versions/Unicode5.1.0/). 1807 13.2. Informative References 1809 [BIG5] Institute for Information Industry of Taiwan, "Computer 1810 Chinese Glyph and Character Code Mapping Table, Technical 1811 Report C-26", 1984. 1813 There are several forms and variations and a closely- 1814 related standard, CNS 11643. See the discussion in 1815 Chapter 3 of Lunde, K., CJKV Information Processing, 1816 O'Reilly & Associates, 1999 1818 [GB18030] "Chinese National Standard GB 18030-2000: Information 1819 Technology -- Chinese ideograms coded character set for 1820 information interchange -- Extension for the basic set.", 1821 2000. 1823 [RFC0810] Feinler, E., Harrenstien, K., Su, Z., and V. White, "DoD 1824 Internet host table specification", RFC 810, March 1982. 1826 [RFC0952] Harrenstien, K., Stahl, M., and E. Feinler, "DoD Internet 1827 host table specification", RFC 952, October 1985. 1829 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 1830 STD 13, RFC 1034, November 1987. 1832 [RFC1035] Mockapetris, P., "Domain names - implementation and 1833 specification", STD 13, RFC 1035, November 1987. 1835 [RFC1123] Braden, R., "Requirements for Internet Hosts - Application 1836 and Support", STD 3, RFC 1123, October 1989. 1838 [RFC2181] Elz, R. and R. Bush, "Clarifications to the DNS 1839 Specification", RFC 2181, July 1997. 1841 [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and 1842 Languages", BCP 18, RFC 2277, January 1998. 1844 [RFC2673] Crawford, M., "Binary Labels in the Domain Name System", 1845 RFC 2673, August 1999. 1847 [RFC2782] Gulbrandsen, A., Vixie, P., and L. Esibov, "A DNS RR for 1848 specifying the location of services (DNS SRV)", RFC 2782, 1849 February 2000. 1851 [RFC3454] Hoffman, P. and M. Blanchet, "Preparation of 1852 Internationalized Strings ("stringprep")", RFC 3454, 1853 December 2002. 1855 [RFC3491] Hoffman, P. and M. Blanchet, "Nameprep: A Stringprep 1856 Profile for Internationalized Domain Names (IDN)", 1857 RFC 3491, March 2003. 1859 [RFC3743] Konishi, K., Huang, K., Qian, H., and Y. Ko, "Joint 1860 Engineering Team (JET) Guidelines for Internationalized 1861 Domain Names (IDN) Registration and Administration for 1862 Chinese, Japanese, and Korean", RFC 3743, April 2004. 1864 [RFC3987] Duerst, M. and M. Suignard, "Internationalized Resource 1865 Identifiers (IRIs)", RFC 3987, January 2005. 1867 [RFC4290] Klensin, J., "Suggested Practices for Registration of 1868 Internationalized Domain Names (IDN)", RFC 4290, 1869 December 2005. 1871 [RFC4690] Klensin, J., Faltstrom, P., Karp, C., and IAB, "Review and 1872 Recommendations for Internationalized Domain Names 1873 (IDNs)", RFC 4690, September 2006. 1875 [RFC4713] Lee, X., Mao, W., Chen, E., Hsu, N., and J. Klensin, 1876 "Registration and Administration Recommendations for 1877 Chinese Domain Names", RFC 4713, October 2006. 1879 [Unicode-Security] 1880 The Unicode Consortium, "Unicode Technical Standard #39: 1881 Unicode Security Mechanisms", August 2008, 1882 . 1884 [Unicode-UAX31] 1885 The Unicode Consortium, "Unicode Standard Annex #31: 1886 Unicode Identifier and Pattern Syntax", March 2008, 1887 . 1889 [Unicode-UTR36] 1890 The Unicode Consortium, "Unicode Technical Report #36: 1891 Unicode Security Considerations", July 2008, 1892 . 1894 Appendix A. Change Log 1896 [[ RFC Editor: Please remove this appendix. ]] 1898 A.1. Changes between Version -00 and Version -01 of 1899 draft-ietf-idnabis-rationale 1901 o Clarified the U-label definition to note that U-labels must 1902 contain at least one non-ASCII character. Also clarified the 1903 relationship among label types. 1905 o Rewrote the discussion of Labels in Registration (Section 7.1.2) 1906 and related text about IDNA-validity (in the "Defs" document as of 1907 -04 of this one) to narrow its focus and remove more general 1908 restrictions. Added a temporary note in line to explain the 1909 situation. 1911 o Changed the "IDNA uses Unicode" statement to focus on 1912 compatibility with IDNA2003 and avoid more general or 1913 controversial assertions. 1915 o Added a discussion of examples to Section 7.1 1917 o Made a number of other small editorial changes and corrections 1918 suggested by Mark Davis. 1920 o Added several more discussion anchors and notes and expanded or 1921 updated some existing ones. 1923 A.2. Version -02 1925 o Trimmed change log, removing information about pre-WG drafts. 1927 o Adjusted discussion of Contextual Rules to match the new location 1928 of the tables and some conceptual material. 1930 o Rewrote the material on preprocessing somewhat. 1932 o Moved the material about relationships with IDNA2003 to be part of 1933 a single section on transitions. 1935 o Removed several placeholders and made editorial changes in 1936 accordance with decisions made at IETF 72 in Dublin and not 1937 disputed on the mailing list. 1939 A.3. Version -03 1941 This special update to the Rationale document is intended to try to 1942 get the discussion of what is normative or not under control. While 1943 the IETF does not normally annotate individual sections of documents 1944 with whether they are normative or not, concerns that we don't know 1945 which is which, claims that some material is normative that would be 1946 problematic if so classified, etc., argue that we should at least be 1947 able to have a clear discussion on the subject. 1949 Two annotations have been applied to sections that might reasonably 1950 be considered normative. One annotation is based on the list of 1951 sections in Mark Davis's note of 29 September (http:// 1952 www.alvestrand.no/pipermail/idna-update/2008-September/002667.html). 1953 The other is based on an elaboration of John Klensin's response on 7 1954 October (http://www.alvestrand.no/pipermail/idna-update/2008-October/ 1955 002691.html). These should just be considered two suggestions to 1956 illuminate and, one hopes, advance the Working Group's discussions. 1958 Some additional editorial changes have been made, but they are 1959 basically trivial. In the editor's judgment, it is not possible to 1960 make significantly more progress with this document until the matter 1961 of document organization is settled. 1963 A.4. Version -04 1965 o Definitional and other normative material moved to new document 1966 (draft-ietf-idnabis-defs). Version -03 annotations removed. 1968 o Material on differences between IDNA2003 and IDNA2003 moved to an 1969 appendix in Protocol. 1971 o Material left over from the origins of this document as a 1972 preliminary proposal has been removed or rewritten. 1974 o Changes made to reflect consensus call results, including removing 1975 several placeholder notes for discussion. 1977 o Added more material, including discussion of historic scripts, to 1978 Section 3.2 on registration policies. 1980 o Added a new section (Section 7.2) to contain specific discussion 1981 of handling of characters that are interpreted differently in 1982 input to IDNA2003 and 2008. 1984 o Some material, including this section/appendix, rearranged. 1986 Author's Address 1988 John C Klensin 1989 1770 Massachusetts Ave, Ste 322 1990 Cambridge, MA 02140 1991 USA 1993 Phone: +1 617 245 1457 1994 Email: john+ietf@jck.com 1996 Full Copyright Statement 1998 Copyright (C) The IETF Trust (2008). 2000 This document is subject to the rights, licenses and restrictions 2001 contained in BCP 78, and except as set forth therein, the authors 2002 retain all their rights. 2004 This document and the information contained herein are provided on an 2005 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 2006 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND 2007 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS 2008 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 2009 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 2010 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 2012 Intellectual Property 2014 The IETF takes no position regarding the validity or scope of any 2015 Intellectual Property Rights or other rights that might be claimed to 2016 pertain to the implementation or use of the technology described in 2017 this document or the extent to which any license under such rights 2018 might or might not be available; nor does it represent that it has 2019 made any independent effort to identify any such rights. Information 2020 on the procedures with respect to rights in RFC documents can be 2021 found in BCP 78 and BCP 79. 2023 Copies of IPR disclosures made to the IETF Secretariat and any 2024 assurances of licenses to be made available, or the result of an 2025 attempt made to obtain a general license or permission for the use of 2026 such proprietary rights by implementers or users of this 2027 specification can be obtained from the IETF on-line IPR repository at 2028 http://www.ietf.org/ipr. 2030 The IETF invites any interested party to bring to its attention any 2031 copyrights, patents or patent applications, or other proprietary 2032 rights that may cover technology that may be required to implement 2033 this standard. Please address the information to the IETF at 2034 ietf-ipr@ietf.org.