idnits 2.17.00 (12 Aug 2021) /tmp/idnits3872/draft-ietf-idnabis-rationale-08.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** The document seems to lack a License Notice according IETF Trust Provisions of 28 Dec 2009, Section 6.b.i or Provisions of 12 Sep 2009 Section 6.b -- however, there's a paragraph with a matching beginning. Boilerplate error? (You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Feb 2009 rather than one of the newer Notices. See https://trustee.ietf.org/license-info/.) Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 944: '... CONTEXTUAL RULE REQUIRED category and...' RFC 2119 keyword, line 1782: '...esented directly in ASCII MUST use the...' RFC 2119 keyword, line 1783: '...sion to A-labels MUST be performed pri...' RFC 2119 keyword, line 2172: '... MUST, SHOULD, or MAY) and small ...' Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document seems to contain a disclaimer for pre-RFC5378 work, and may have content which was first submitted before 10 November 2008. The disclaimer is necessary when there are original authors that you have been unable to contact, or if some do not wish to grant the BCP78 rights to the IETF Trust. If you are able to get all authors (current and original) to grant those rights, you can and should remove the disclaimer; otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (March 6, 2009) is 4823 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'ASCII' is defined on line 1923, but no explicit reference was found in the text == Unused Reference: 'RFC1034' is defined on line 2000, but no explicit reference was found in the text == Unused Reference: 'RFC1035' is defined on line 2003, but no explicit reference was found in the text == Unused Reference: 'RFC2673' is defined on line 2025, but no explicit reference was found in the text == Unused Reference: 'Unicode-UTR36' is defined on line 2070, but no explicit reference was found in the text == Outdated reference: draft-ietf-idnabis-defs has been published as RFC 5890 == Outdated reference: draft-ietf-idnabis-tables has been published as RFC 5892 ** Obsolete normative reference: RFC 3490 (Obsoleted by RFC 5890, RFC 5891) -- Obsolete informational reference (is this intentional?): RFC 810 (Obsoleted by RFC 952) -- Obsolete informational reference (is this intentional?): RFC 2535 (Obsoleted by RFC 4033, RFC 4034, RFC 4035) -- Obsolete informational reference (is this intentional?): RFC 2671 (Obsoleted by RFC 6891) -- Obsolete informational reference (is this intentional?): RFC 2673 (Obsoleted by RFC 6891) -- Obsolete informational reference (is this intentional?): RFC 3454 (Obsoleted by RFC 7564) -- Obsolete informational reference (is this intentional?): RFC 3491 (Obsoleted by RFC 5891) Summary: 3 errors (**), 0 flaws (~~), 8 warnings (==), 8 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group J. Klensin 3 Internet-Draft March 6, 2009 4 Intended status: Informational 5 Expires: September 7, 2009 7 Internationalized Domain Names for Applications (IDNA): Background, 8 Explanation, and Rationale 9 draft-ietf-idnabis-rationale-08.txt 11 Status of this Memo 13 This Internet-Draft is submitted to IETF in full conformance with the 14 provisions of BCP 78 and BCP 79. This document may contain material 15 from IETF Documents or IETF Contributions published or made publicly 16 available before November 10, 2008. The person(s) controlling the 17 copyright in some of this material may not have granted the IETF 18 Trust the right to allow modifications of such material outside the 19 IETF Standards Process. Without obtaining an adequate license from 20 the person(s) controlling the copyright in such materials, this 21 document may not be modified outside the IETF Standards Process, and 22 derivative works of it may not be created outside the IETF Standards 23 Process, except to format it for publication as an RFC or to 24 translate it into languages other than English. 26 Internet-Drafts are working documents of the Internet Engineering 27 Task Force (IETF), its areas, and its working groups. Note that 28 other groups may also distribute working documents as Internet- 29 Drafts. 31 Internet-Drafts are draft documents valid for a maximum of six months 32 and may be updated, replaced, or obsoleted by other documents at any 33 time. It is inappropriate to use Internet-Drafts as reference 34 material or to cite them other than as "work in progress." 36 The list of current Internet-Drafts can be accessed at 37 http://www.ietf.org/ietf/1id-abstracts.txt. 39 The list of Internet-Draft Shadow Directories can be accessed at 40 http://www.ietf.org/shadow.html. 42 This Internet-Draft will expire on September 7, 2009. 44 Copyright Notice 46 Copyright (c) 2009 IETF Trust and the persons identified as the 47 document authors. All rights reserved. 49 This document is subject to BCP 78 and the IETF Trust's Legal 50 Provisions Relating to IETF Documents in effect on the date of 51 publication of this document (http://trustee.ietf.org/license-info). 52 Please review these documents carefully, as they describe your rights 53 and restrictions with respect to this document. 55 Abstract 57 Several years have passed since the original protocol for 58 Internationalized Domain Names (IDNs) was completed and deployed. 59 During that time, a number of issues have arisen, including the need 60 to update the system to deal with newer versions of Unicode. Some of 61 these issues require tuning of the existing protocols and the tables 62 on which they depend. This document provides an overview of a 63 revised system and provides explanatory material for its components. 65 Table of Contents 67 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 68 1.1. Context and Overview . . . . . . . . . . . . . . . . . . . 4 69 1.2. Discussion Forum . . . . . . . . . . . . . . . . . . . . . 4 70 1.3. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 71 1.3.1. Documents and Standards . . . . . . . . . . . . . . . 5 72 1.3.2. DNS "Name" Terminology . . . . . . . . . . . . . . . . 5 73 1.3.3. New Terminology and Restrictions . . . . . . . . . . . 6 74 1.4. Objectives . . . . . . . . . . . . . . . . . . . . . . . . 6 75 1.5. Applicability and Function of IDNA . . . . . . . . . . . . 7 76 1.6. Comprehensibility of IDNA Mechanisms and Processing . . . 8 77 2. Processing in IDNA2008 . . . . . . . . . . . . . . . . . . . . 9 78 3. Permitted Characters: An Inclusion List . . . . . . . . . . . 10 79 3.1. A Tiered Model of Permitted Characters and Labels . . . . 10 80 3.1.1. PROTOCOL-VALID . . . . . . . . . . . . . . . . . . . . 11 81 3.1.2. Characters Valid Only in Context With Others . . . . . 11 82 3.1.2.2. Rules and Their Application . . . . . . . . . . . 12 83 3.1.3. DISALLOWED . . . . . . . . . . . . . . . . . . . . . . 12 84 3.1.4. UNASSIGNED . . . . . . . . . . . . . . . . . . . . . . 13 85 3.2. Registration Policy . . . . . . . . . . . . . . . . . . . 14 86 3.3. Layered Restrictions: Tables, Context, Registration, 87 Applications . . . . . . . . . . . . . . . . . . . . . . . 14 88 4. Issues that Constrain Possible Solutions . . . . . . . . . . . 15 89 4.1. Display and Network Order . . . . . . . . . . . . . . . . 15 90 4.2. Entry and Display in Applications . . . . . . . . . . . . 16 91 4.3. Linguistic Expectations: Ligatures, Digraphs, and 92 Alternate Character Forms . . . . . . . . . . . . . . . . 17 93 4.4. Case Mapping and Related Issues . . . . . . . . . . . . . 20 94 4.5. Right to Left Text . . . . . . . . . . . . . . . . . . . . 20 95 5. IDNs and the Robustness Principle . . . . . . . . . . . . . . 21 96 6. Front-end and User Interface Processing for Lookup . . . . . . 22 97 7. Migration from IDNA2003 and Unicode Version Synchronization . 25 98 7.1. Design Criteria . . . . . . . . . . . . . . . . . . . . . 25 99 7.1.1. General IDNA Validity Criteria . . . . . . . . . . . . 25 100 7.1.2. Labels in Registration . . . . . . . . . . . . . . . . 26 101 7.1.3. Labels in Lookup . . . . . . . . . . . . . . . . . . . 27 102 7.2. Changes in Character Interpretations . . . . . . . . . . . 28 103 7.3. More Flexibility in User Agents . . . . . . . . . . . . . 30 104 7.4. The Question of Prefix Changes . . . . . . . . . . . . . . 31 105 7.4.1. Conditions Requiring a Prefix Change . . . . . . . . . 32 106 7.4.2. Conditions Not Requiring a Prefix Change . . . . . . . 32 107 7.4.3. Implications of Prefix Changes . . . . . . . . . . . . 33 108 7.5. Stringprep Changes and Compatibility . . . . . . . . . . . 33 109 7.6. The Symbol Question . . . . . . . . . . . . . . . . . . . 34 110 7.7. Migration Between Unicode Versions: Unassigned Code 111 Points . . . . . . . . . . . . . . . . . . . . . . . . . . 35 112 7.8. Other Compatibility Issues . . . . . . . . . . . . . . . . 37 113 8. Name Server Considerations . . . . . . . . . . . . . . . . . . 37 114 8.1. Processing Non-ASCII Strings . . . . . . . . . . . . . . . 37 115 8.2. DNSSEC Authentication of IDN Domain Names . . . . . . . . 38 116 8.3. Root and other DNS Server Considerations . . . . . . . . . 38 117 9. Internationalization Considerations . . . . . . . . . . . . . 39 118 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 39 119 10.1. IDNA Character Registry . . . . . . . . . . . . . . . . . 39 120 10.2. IDNA Context Registry . . . . . . . . . . . . . . . . . . 39 121 10.3. IANA Repository of IDN Practices of TLDs . . . . . . . . . 39 122 11. Security Considerations . . . . . . . . . . . . . . . . . . . 40 123 11.1. General Security Issues with IDNA . . . . . . . . . . . . 40 124 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 40 125 13. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 41 126 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 41 127 14.1. Normative References . . . . . . . . . . . . . . . . . . . 41 128 14.2. Informative References . . . . . . . . . . . . . . . . . . 42 129 Appendix A. Change Log . . . . . . . . . . . . . . . . . . . . . 44 130 A.1. Changes between Version -00 and Version -01 of 131 draft-ietf-idnabis-rationale . . . . . . . . . . . . . . . 44 132 A.2. Version -02 . . . . . . . . . . . . . . . . . . . . . . . 45 133 A.3. Version -03 . . . . . . . . . . . . . . . . . . . . . . . 45 134 A.4. Version -04 . . . . . . . . . . . . . . . . . . . . . . . 46 135 A.5. Version -05 . . . . . . . . . . . . . . . . . . . . . . . 46 136 A.6. Version -06 . . . . . . . . . . . . . . . . . . . . . . . 46 137 A.7. Version -07 . . . . . . . . . . . . . . . . . . . . . . . 47 138 A.8. Version -08 . . . . . . . . . . . . . . . . . . . . . . . 47 139 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 47 141 1. Introduction 143 1.1. Context and Overview 145 The original standards for Internationalized Domain Names (IDNs) were 146 completed and deployed starting in 2003. Those standards are known 147 as Internationalized Domain Names in Applications (IDNA), taken from 148 the name of the highest level standard within the group, RFC 3490 149 [RFC3490]. After those standards were deployed, a number of issues 150 arose that led to a call for a new version of the IDNA protocol and 151 the associated tables, including a subset of those described in a 152 recent IAB report [RFC4690] and the need to update the system to deal 153 with newer versions of Unicode. This document further explains the 154 issues that have been encountered when they are important to 155 understanding of the revised protocols. It also provides an overview 156 of the new IDNA model and explanatory material for it. Additional 157 explanatory material for the specific components of the proposals 158 appears with the associated documents. 160 This document and the associated ones are written from the 161 perspective of an IDNA-aware user, application, or implementation. 162 While they may reiterate fundamental DNS rules and requirements for 163 the convenience of the reader, they make no attempt to be 164 comprehensive about DNS principles and should not be considered as a 165 substitute for a thorough understanding of the DNS protocols and 166 specifications. 168 A good deal of the background material that appeared in RFC 3490 169 [RFC3490] has been removed from this update. That material is either 170 of historical interest only or has been covered from a more recent 171 perspective in RFC 4690 [RFC4690]. 173 This document is not normative. The information it provides is 174 intended to make the rules, tables, and protocol easier to understand 175 and to provide overview information and suggestions for zone 176 administrators and others who need to make policy, deployment, and 177 similar decisions about IDNs. 179 1.2. Discussion Forum 181 [[ RFC Editor: please remove this section. ]] 183 IDNA2008 is being discussed in the IETF "idnabis" Working Group and 184 on the mailing list idna-update@alvestrand.no 186 1.3. Terminology 188 Terminology that is critical for understanding this document and the 189 rest of the documents that make up IDNA2008, appears in 190 [IDNA2008-Defs]. That document also contains roadmap to the IDNA2008 191 document collection. No attempt should be made to understand this 192 document without the definitions and concepts that appear there. 194 1.3.1. Documents and Standards 196 This document uses the term "IDNA2003" to refer to the set of 197 standards that make up and support the version of IDNA published in 198 2003, i.e., those commonly known as the IDNA base specification 199 [RFC3490], Nameprep [RFC3491], Punycode [RFC3492], and Stringprep 200 [RFC3454]. In this document, those names are used to refer, 201 conceptually, to the individual documents, with the base IDNA 202 specification called just "IDNA". 204 The term "IDNA2008" is used to refer to a new version of IDNA as 205 described in this document and in the documents described in the 206 document listing of [IDNA2008-Defs]. IDNA2008 is not dependent on 207 any of the IDNA2003 specifications other than the one for Punycode 208 encoding. References to "these specifications" or "these documents" 209 are to the entire IDNA2008 set. 211 1.3.2. DNS "Name" Terminology 213 These documents depart from historical DNS terminology and usage in 214 one important respect. Over the years, the community has talked very 215 casually about "names" in the DNS, beginning with calling it "the 216 domain name system". That terminology is fine in the very precise 217 sense that the identifiers of the DNS do provide names for objects 218 and addresses. But, in the context of IDNs, the term has introduced 219 some confusion, confusion that has increased further as people have 220 begun to speak of DNS labels in terms of the words or phrases of 221 various natural languages. 223 Historically, many, perhaps most, of the "names" in the DNS have been 224 mnemonics to identify some particular concept, object, or 225 organization. They are typically derived from, or rooted in, some 226 language because most people think in language-based ways. But, 227 because they are mnemonics, they need not obey the orthographic 228 conventions of any language: it is not a requirement that it be 229 possible for them to be "words". 231 This distinction is important because the reasonable goal of an IDN 232 effort is not to be able to write the great Klingon (or language of 233 one's choice) novel in DNS labels but to be able to form a usefully 234 broad range of mnemonics in ways that are as natural as possible in a 235 very broad range of scripts. 237 1.3.3. New Terminology and Restrictions 239 These documents introduce new terminology, and precise definitions, 240 for the terms "U-label", "A-Label", LDH-label (to which all valid 241 pre-IDNA host names conformed), Reserved-LDH-label (R-LDH-label), XN- 242 label, Fake-A-Label, and Non-Reserved-LDH-label (NR-LDH-label). 244 In addition, the term "putative label" has been adopted to refer to a 245 label that may appear to meet certain definitional constraints but 246 has not yet been sufficiently tested for validity. 248 These definitions are illustrated in Figure 1 of the Definitions 249 Document [IDNA2008-Defs]. R-LDH-labels contain "--" in the third and 250 fourth character from the beginning of the label. In IDNA-aware 251 applications, only a subset of these reserved labels is permitted to 252 be used, namely the A-label subset. A-labels are a subset of the 253 R-LDH-labels that begin with the case-insensitive (?) string "xn--". 254 Labels that bear this prefix but which are not otherwise valid fall 255 into the "Fake-A-label" category. The non-reserved labels (NR-LDH- 256 labels) are implicitly valid since they do not trigger any 257 resemblance to IDNA-landr NR-LDH-labels. 259 The creation of the Reserved-LDH category is required for three 260 reasons: 262 o to prevent confusion with pre-IDNA coding forms; 264 o to permit future extensions that would require changing the 265 prefix, no matter how unlikely those might be (see Section 7.4); 266 and 268 o to reduce the opportunities for attacks via the Punycode encoding 269 algorithm itself. 271 1.4. Objectives 273 The intent of the IDNA revision effort, and hence of this document 274 and the associated ones, is to increase the usability and 275 effectiveness of internationalized domain names (IDNs) while 276 preserving or strengthening the integrity of references that use 277 them. The original "hostname" character definitions (see, e.g., 278 [RFC0810]) struck a balance between the creation of useful mnemonics 279 and the introduction of parsing problems or general confusion in the 280 contexts in which domain names are used. The objective of IDNA2008 281 is to preserve that balance while expanding the character repertoire 282 to include extended versions of Roman-derived scripts and scripts 283 that are not Roman in origin. No work of this sort is able to 284 completely eliminate sources of visual or textual confusion: such 285 confusion is possible even under the original host naming rules where 286 only ASCII characters were permitted. However, through the 287 application of different techniques at different points (see 288 Section 3.3), it should be possible to keep problems to an acceptable 289 minimum. One consequence of this general objective is that the 290 desire of some user or marketing community to use a particular string 291 --whether the reason is to try to write sentences of particular 292 languages in the DNS, to express a facsimile of the symbol for a 293 brand, or for some other purpose-- is not a primary goal within the 294 context of applications in the domain name space. 296 1.5. Applicability and Function of IDNA 298 The IDNA specification solves the problem of extending the repertoire 299 of characters that can be used in domain names to include a large 300 subset of the Unicode repertoire. 302 IDNA does not extend the service offered by DNS to the applications. 303 Instead, the applications (and, by implication, the users) continue 304 to see an exact-match lookup service. Either there is a single 305 exactly-matching name or there is no match. This model has served 306 the existing applications well, but it requires, with or without 307 internationalized domain names, that users know the exact spelling of 308 the domain names that are to be typed into applications such as web 309 browsers and mail user agents. The introduction of the larger 310 repertoire of characters potentially makes the set of misspellings 311 larger, especially given that in some cases the same appearance, for 312 example on a business card, might visually match several Unicode code 313 points or several sequences of code points. 315 The IDNA standard does not require any applications to conform to it, 316 nor does it retroactively change those applications. An application 317 can elect to use IDNA in order to support IDN while maintaining 318 interoperability with existing infrastructure. If an application 319 wants to use non-ASCII characters in domain names, IDNA is the only 320 currently-defined option. Adding IDNA support to an existing 321 application entails changes to the application only, and leaves room 322 for flexibility in front-end processing and more specifically in the 323 user interface (see Section 6). 325 A great deal of the discussion of IDN solutions has focused on 326 transition issues and how IDNs will work in a world where not all of 327 the components have been updated. Proposals that were not chosen by 328 the original IDN Working Group would have depended on updating of 329 user applications, DNS resolvers, and DNS servers in order for a user 330 to apply an internationalized domain name in any form or coding 331 acceptable under that method. While processing must be performed 332 prior to or after access to the DNS, IDNA requires no changes to the 333 DNS protocol or any DNS servers or the resolvers on user's computers. 335 IDNA allows the graceful introduction of IDNs not only by avoiding 336 upgrades to existing infrastructure (such as DNS servers and mail 337 transport agents), but also by allowing some rudimentary use of IDNs 338 in applications by using the ASCII-encoded representation of the 339 labels containing non-ASCII characters. While such names are user- 340 unfriendly to read and type, and hence not optimal for user input, 341 they can be used as a last resort to allow rudimentary IDN usage. 342 For example, they might be the best choice for display if it were 343 known that relevant fonts were not available on the user's computer. 344 In order to allow user-friendly input and output of the IDNs and 345 acceptance of some characters as equivalent to those to be processed 346 according to the protocol, the applications need to be modified to 347 conform to this specification. 349 This version of IDNA uses the Unicode character repertoire, for 350 continuity with the original version of IDNA. 352 1.6. Comprehensibility of IDNA Mechanisms and Processing 354 One of the major goals of this work is to improve the general 355 understanding of how IDNA works and what characters are permitted and 356 what happens to them. Comprehensibility and predictability to users 357 and registrants are themselves important motivations and design goals 358 for this effort. The effort includes some new terminology and a 359 revised and extended model, both covered in this section, and some 360 more specific protocol, processing, and table modifications. Details 361 of the latter appear in other documents (see [IDNA2008-Defs]). 363 Several issues are inherent in the application of IDNs and, indeed, 364 almost any other system that tries to handle international characters 365 and concepts. They range from the apparently trivial --e.g., one 366 cannot display a character for which one does not have a font 367 available locally-- to the more complex and subtle. Many people have 368 observed that internationalization is just a tool to enable effective 369 localization while permitting some global uniformity. Issues of 370 display, of exactly how various strings and characters are entered, 371 and so on are inherently issues about localization and user interface 372 design. 374 A protocol such as IDNA can only assume that such operations as data 375 entry and reconciliation of differences in character forms are 376 possible. It may make some recommendations about how display might 377 work when characters and fonts are not available, but they can only 378 be general recommendations and, because display functions are rarely 379 controlled by the types of applications that would call upon IDNA, 380 will rarely be very effective. 382 However, shifting responsibility for character mapping and other 383 adjustments from the protocol (where it was located in IDNA2003) to 384 the user interface or processing before invoking IDNA raises issues 385 about both what that processing should do and about compatibility for 386 references prepared in an IDNA2003 context. Those issues are 387 discussed in Section 6. 389 Operations for converting between local character sets and normalized 390 Unicode are part of this general set of user interface issues. The 391 conversion is obviously not required at all in a Unicode-native 392 system that maintains all strings in Normalization Form C (NFC). 393 (See [Unicode-UAX15] for precise definitions of NFC and NFKC if 394 needed.) It may, however, involve some complexity in a system that 395 is not Unicode-native, especially if the elements of the local 396 character set do not map exactly and unambiguously into Unicode 397 characters or do so in a way that is not completely stable over time. 398 Perhaps more important, if a label being converted to a local 399 character set contains Unicode characters that have no correspondence 400 in that character set, the application may have to apply special, 401 locally-appropriate, methods to avoid or reduce loss of information. 403 Depending on the system involved, the major difficulty may not lie in 404 the mapping but in accurately identifying the incoming character set 405 and then applying the correct conversion routine. If a local 406 operating system uses one of the ISO 8859 character sets or an 407 extensive national or industrial system such as GB18030 [GB18030] or 408 BIG5 [BIG5], one must correctly identify the character set in use 409 before converting to Unicode even though those character coding 410 systems are substantially or completely Unicode-compatible (i.e., all 411 of the code points in them have an exact and unique mapping to 412 Unicode code points). It may be even more difficult when the 413 character coding system in local use is based on conceptually 414 different assumptions than those used by Unicode about, e.g., font 415 encodings used for publications in some Indic scripts. Those 416 differences may not easily yield unambiguous conversions or 417 interpretations even if each coding system is internally consistent 418 and adequate to represent the local language and script. 420 2. Processing in IDNA2008 422 These specifications separate Domain Name Registration and Lookup in 423 the protocol specification. Doing so reflects current practice in 424 which per-registry restrictions and special processing are applied at 425 registration time but not during lookup. Even more important in the 426 longer term, it facilitates incremental addition of permitted 427 character groups to avoid freezing on one particular version of 428 Unicode. 430 The actual registration and lookup protocols for IDNA2008 are 431 specified in [IDNA2008-Protocol]. 433 3. Permitted Characters: An Inclusion List 435 This section provides an overview of the model used to establish the 436 algorithm and character lists of [IDNA2008-Tables] and describes the 437 names and applicability of the categories used there. Note that the 438 inclusion of a character in the first category group (Section 3.1.1) 439 does not imply that it can be used indiscriminately; some characters 440 are associated with contextual rules that must be applied as well. 442 The information given in this section is provided to make the rules, 443 tables, and protocol easier to understand. The normative generating 444 rules that correspond to this informal discussion appear in 445 [IDNA2008-Tables] and the rules that actually determine what labels 446 can be registered or looked up are in [IDNA2008-Protocol]. 448 3.1. A Tiered Model of Permitted Characters and Labels 450 Moving to an inclusion model requires respecifying the list of 451 characters that are permitted in IDNs. In IDNA2003, the role and 452 utility of characters are independent of context and fixed forever 453 (or until the standard is replaced). Making completely context- 454 independent rules globally has proven impractical because some 455 characters, especially those that are called "Join_Controls" in 456 Unicode, are needed to make reasonable use of some scripts but have 457 no visible effect(s) in others. IDNA2003 prohibited those types of 458 characters entirely. But the restrictions led to a consensus that 459 under some conditions, these "joiner" characters were legitimately 460 needed to allow useful mnemonics for some languages and scripts. The 461 requirement to support those characters but limit their use to very 462 specific contexts was reinforced by the observation that handling of 463 particular characters across the languages that use a script, or the 464 use of similar or identical-looking characters in different scripts, 465 is more complex than many people believed it was several years ago. 467 Independently of the characters chosen (see next subsection), the 468 approach is to divide the characters that appear in Unicode into 469 three categories: 471 3.1.1. PROTOCOL-VALID 473 Characters identified as "PROTOCOL-VALID" (often abbreviated 474 "PVALID") are, in general, permitted by IDNA for all uses in IDNs. 475 Their use may be restricted by rules about the context in which they 476 appear or by other rules that apply to the entire label in which they 477 are to be embedded. For example, any label that contains a character 478 in this category that has a "right-to-left" property must be used in 479 context with the "Bidi" rules (see [IDNA2008-Bidi]). 481 The term "PROTOCOL-VALID" is used to stress the fact that the 482 presence of a character in this category does not imply that a given 483 registry need accept registrations containing any of the characters 484 in the category. Registries are still expected to apply judgment 485 about labels they will accept and to maintain rules consistent with 486 those judgments (see [IDNA2008-Protocol] and Section 3.3). 488 Characters that are placed in the "PROTOCOL-VALID" category are 489 expected to never be removed from it or reclassified. While 490 theoretically characters could be removed from Unicode, such removal 491 would be inconsistent with the Unicode stability principles (see 492 [Unicode51], Appendix F) and hence should never occur. 494 3.1.2. Characters Valid Only in Context With Others 496 Some characters may be unsuitable for general use in IDNs but 497 necessary for the plausible support of some scripts. The two most 498 commonly-cited examples are the zero-width joiner and non-joiner 499 characters (ZWJ, U+200D and ZWNJ, U+200C), but provisions for 500 unambiguous labels may require that other characters be restricted to 501 particular contexts. For example, the ASCII hyphen is not permitted 502 to start or end a label, whether that label contains non-ASCII 503 characters or not. 505 3.1.2.1. Contextual Restrictions 507 These characters must not appear in IDNs without additional 508 restrictions, typically because they have no visible consequences in 509 most scripts but affect format or presentation in a few others or 510 because they are combining characters that are safe for use only in 511 conjunction with particular characters or scripts. In order to 512 permit them to be used at all, they are specially identified as 513 "CONTEXTUAL RULE REQUIRED" and, when adequately understood, 514 associated with a rule. In addition, the rule will define whether it 515 is to be applied on lookup as well as registration. A distinction is 516 made between characters that indicate or prohibit joining (known as 517 "CONTEXT-JOINER" or "CONTEXTJ") and other characters requiring 518 contextual treatment ("CONTEXT-OTHER" or "CONTEXTO"). Only the 519 former require full testing at lookup time. 521 It is important to note that these contextual rules cannot prevent 522 all uses of the relevant characters that might be confusing or 523 problematic. What they are expected do is to confine applicability 524 of the characters to scripts (and narrower contexts) where zone 525 administrators are knowledgeable enough about the use of those 526 characters to be prepared to deal with them appropriately. For 527 example, a registry dealing with an Indic script that requires ZWJ 528 and/or ZWNJ as part of the writing system is expected to understand 529 where the characters have visible effect and where they do not and to 530 make registration rules accordingly. By contrast, a registry dealing 531 with Latin or Cyrillic script might not be actively aware that the 532 characters exist, much less about the consequences of embedding them 533 in labels drawn from those scripts. 535 3.1.2.2. Rules and Their Application 537 The actual rules may be DEFINED or NULL. If present, they may have 538 values of "True" (character may be used in any position in any 539 label), "False" (character may not be used in any label), or may be a 540 set of procedural rules that specify the context in which the 541 character is permitted. 543 Examples of descriptions of typical rules, stated informally and in 544 English, include "Must follow a character from Script XYZ", "Must 545 occur only if the entire label is in Script ABC", "Must occur only if 546 the previous and subsequent characters have the DFG property". 548 Because it is easier to identify these characters than to know that 549 they are actually needed in IDNs or how to establish exactly the 550 right rules for each one, a rule may have a null value in a given 551 version of the tables. Characters associated with null rules are not 552 permitted to appear in putative labels for either registration or 553 lookup. Of course, a later version of the tables might contain a 554 non-null rule. 556 The description of the syntax of the rules, and the rules themselves, 557 appears in [IDNA2008-Tables]. [[anchor11: ??? Section number would 558 be good here.]] 560 3.1.3. DISALLOWED 562 Some characters are inappropriate for use in IDNs and are thus 563 excluded for both registration and lookup (i.e., IDNA-conforming 564 applications performing name lookup should verify that these 565 characters are absent; if they are present, the label strings should 566 be rejected rather than converted to A-labels and looked up. Some of 567 these characters are problematic for use in IDNs (such as the 568 FRACTION SLASH character, U+2044), while some of them (such as the 569 various HEART symbols, e.g., U+2665, U+2661, and U+2765, see 570 Section 7.6) simply fall outside the conventions for typical 571 identifiers (basically letters and numbers). 573 Of course, this category would include code points that had been 574 removed entirely from Unicode should such removals ever occur. 576 Characters that are placed in the "DISALLOWED" category are expected 577 to never be removed from it or reclassified. If a character is 578 classified as "DISALLOWED" in error and the error is sufficiently 579 problematic, the only recourse would be either to introduce a new 580 code point into Unicode and classify it as "PROTOCOL-VALID" or for 581 the IETF to accept the considerable costs of an incompatible change 582 and replace the relevant RFC with one containing appropriate 583 exceptions. 585 There is provision for exception cases but, in general, characters 586 are placed into "DISALLOWED" if they fall into one or more of the 587 following groups: 589 o The character is a compatibility equivalent for another character. 590 In slightly more precise Unicode terms, application of 591 normalization method NFKC to the character yields some other 592 character. 594 o The character is an upper-case form or some other form that is 595 mapped to another character by Unicode casefolding. 597 o The character is a symbol or punctuation form or, more generally, 598 something that is not a letter, digit, or a mark that is used to 599 form a letter or digit. 601 3.1.4. UNASSIGNED 603 For convenience in processing and table-building, code points that do 604 not have assigned values in a given version of Unicode are treated as 605 belonging to a special UNASSIGNED category. Such code points are 606 prohibited in labels to be registered or looked up. The category 607 differs from DISALLOWED in that code points are moved out of it by 608 the simple expedient of being assigned in a later version of Unicode 609 (at which point, they are classified into one of the other categories 610 as appropriate). 612 The rationale for restricting the processing of UNASSIGNED characters 613 is simply that if such characters were permitted to be looked up, for 614 example, and were later assigned, but subject to some set of 615 contextual rules, un-updated instances of IDNA-aware software might 616 permit lookup of labels containing the previously-unassigned 617 characters while updated versions of IDNA-aware software might 618 restrict their use in lookup, depending on the contextual rules. It 619 should be clear that under no circumstance should an UNASSIGNED 620 character be permitted in a label to be registered as part of a 621 domain name. 623 3.2. Registration Policy 625 While these recommendations cannot and should not define registry 626 policies, registries should develop and apply additional restrictions 627 as needed to reduce confusion and other problems. For example, it is 628 generally believed that labels containing characters from more than 629 one script are a bad practice although there may be some important 630 exceptions to that principle. Some registries may choose to restrict 631 registrations to characters drawn from a very small number of 632 scripts. For many scripts, the use of variant techniques such as 633 those as described in RFC 3843 [RFC3743] and RFC 4290 [RFC4290], and 634 illustrated for Chinese by the tables described in RFC 4713 [RFC4713] 635 may be helpful in reducing problems that might be perceived by users. 637 In general, users will benefit if registries only permit characters 638 from scripts that are well-understood by the registry or its 639 advisers. If a registry decides to reduce opportunities for 640 confusion by constructing policies that disallow characters used in 641 historic writing systems or characters whose use is restricted to 642 specialized, highly technical contexts, some relevant information may 643 be found in Section 2.4 "Specific Character Adjustments", Table 4 644 "Candidate Characters for Exclusion from Identifiers" of 645 [Unicode-UAX31] and Section 3.1. "General Security Profile for 646 Identifiers" in [Unicode-Security]. 648 It is worth stressing that these principles of policy development and 649 application apply at all levels of the DNS, not only, e.g., TLD or 650 SLD registrations and that even a trivial, "anything permitted that 651 is valid under the protocol" policy is helpful in that it helps users 652 and application developers know what to expect. 654 3.3. Layered Restrictions: Tables, Context, Registration, Applications 656 The essence of the character rules in IDNA2008 is based on the 657 realization that there is no single magic bullet for any of the 658 issues associated with a multiscript DNS. Instead, the 659 specifications define a variety of approaches that, together, 660 constitute multiple lines of defense against ambiguity in identifiers 661 and loss of referential integrity. The actual character tables are 662 the first mechanism, protocol rules about how those characters are 663 applied or restricted in context are the second, and those two in 664 combination constitute the limits of what can be done by a protocol 665 alone. As discussed in the previous section (Section 3.2), 666 registries are expected to restrict what they permit to be 667 registered, devising and using rules that are designed to optimize 668 the balance between confusion and risk on the one hand and maximum 669 expressiveness in mnemonics on the other. 671 In addition, there is an important role for user agents in warning 672 against label forms that appear problematic given their knowledge of 673 local contexts and conventions. Of course, no approach based on 674 naming or identifiers alone can protect against all threats. 676 4. Issues that Constrain Possible Solutions 678 4.1. Display and Network Order 680 The correct treatment of domain names requires a clear distinction 681 between Network Order (the order in which the code points are sent in 682 protocols) and Display Order (the order in which the code points are 683 displayed on a screen or paper). The order of labels in a domain 684 name that contains characters that are normally written right to left 685 is discussed in [IDNA2008-Bidi]. In particular, there are questions 686 about the order in which labels are displayed if left to right and 687 right to left labels are adjacent to each other, especially if there 688 are also multiple consecutive appearances of one of the types. The 689 decision about the display order is ultimately under the control of 690 user agents --including web browsers, mail clients, and the like-- 691 which may be highly localized. Even when formats are specified by 692 protocols, the full composition of an Internationalized Resource 693 Identifier (IRI) [RFC3987] or Internationalized Email address 694 contains elements other than the domain name. For example, IRIs 695 contain protocol identifiers and field delimiter syntax such as 696 "http://" or "mailto:" while email addresses contain the "@" to 697 separate local parts from domain names. User agents are not required 698 to use those protocol-based forms directly but often do so. While 699 display, parsing, and processing within a label is specified by the 700 normative documents in the IDNA2008 collection, the relationship 701 between fully-qualified domain names and internationalized labels is 702 unchanged from the base DNS specifications. Comments in this 703 document about such full domain names are explanatory or examples of 704 what might be done and must not be considered normative. 706 Questions remain about protocol constraints implying that the overall 707 direction of these strings will always be left to right (or right to 708 left) for an IRI or email address, or if they even should conform to 709 such rules. These questions also have several possible answers. 711 Should a domain name abc.def, in which both labels are represented in 712 scripts that are written right to left, be displayed as fed.cba or 713 cba.fed? An IRI for clear text web access would, in network order, 714 begin with "http://" and the characters will appear as 715 "http://abc.def" -- but what does this suggest about the display 716 order? When entering a URI to many browsers, it may be possible to 717 provide only the domain name and leave the "http://" to be filled in 718 by default, assuming no tail (an approach that does not work for 719 protocols other than HTTP or whatever is chosen as the default). The 720 natural display order for the typed domain name on a right to left 721 system is fed.cba. Does this change if a protocol identifier, tail, 722 and the corresponding delimiters are specified? 724 While logic, precedent, and reality suggest that these are questions 725 for user interface design, not IETF protocol specifications, 726 experience in the 1980s and 1990s with mixing systems in which domain 727 name labels were read in network order (left to right) and those in 728 which those labels were read right to left would predict a great deal 729 of confusion, and heuristics that sometimes fail, if each 730 implementation of each application makes its own decisions on these 731 issues. 733 Any version of IDNA, including the current one, must be written in 734 terms of the network (transmission on the wire) order of characters 735 in labels and for the labels in complete (fully-qualified) domain 736 names and must be quite precise about those relationships. While 737 some strong suggestions about display order would be desirable to 738 reduce the chances for inconsistent transcription of domain names 739 from printed form, such suggestions are beyond the scope of these 740 specifications. 742 4.2. Entry and Display in Applications 744 Applications can accept domain names using any character set or sets 745 desired by the application developer, specified by the operating 746 system, or dictated by other constraints, and can display domain 747 names in any character set or character coding system. That is, the 748 IDNA protocol does not affect the interface between users and 749 applications. 751 An IDNA-aware application can accept and display internationalized 752 domain names in two formats: the internationalized character set(s) 753 supported by the application (i.e., an appropriate local 754 representation of a U-label), and as an A-label. Applications may 755 allow the display of A-labels, but are encouraged to not do so except 756 as an interface for special purposes, possibly for debugging, or to 757 cope with display limitations. In general, they should allow, but 758 not encourage, user input of that label form. A-labels are opaque 759 and ugly and malicious variations on them are not easily detected by 760 users. Where possible, they should thus only be exposed to users and 761 in contexts in which they are absolutely needed. Because IDN labels 762 can be rendered either as A-labels or U-labels, the application may 763 reasonably have an option for the user to select the preferred method 764 of display; if it does, rendering the U-label should normally be the 765 default. 767 Domain names are often stored and transported in many places. For 768 example, they are part of documents such as mail messages and web 769 pages. They are transported in many parts of many protocols, such as 770 both the control commands of SMTP and associated the message body 771 parts, and in the headers and the body content in HTTP. It is 772 important to remember that domain names appear both in domain name 773 slots and in the content that is passed over protocols. 775 In protocols and document formats that define how to handle 776 specification or negotiation of charsets, labels can be encoded in 777 any charset allowed by the protocol or document format. If a 778 protocol or document format only allows one charset, the labels must 779 be given in that charset. Of course, not all charsets can properly 780 represent all labels. If a U-label cannot be displayed in its 781 entirety, the only choice (without loss of information) may be to 782 display the A-label. 784 In any place where a protocol or document format allows transmission 785 of the characters in internationalized labels, labels should be 786 transmitted using whatever character encoding and escape mechanism 787 the protocol or document format uses at that place. This provision 788 is intended to prevent situations in which, e.g., UTF-8 domain names 789 appear embedded in text that is otherwise in some other character 790 coding. 792 All protocols that use domain name slots (See Section 2.3.1.6 793 [[anchor14: ?? Verify this]] in [IDNA2008-Defs]) already have the 794 capacity for handling domain names in the ASCII charset. Thus, 795 A-labels can inherently be handled by those protocols. 797 4.3. Linguistic Expectations: Ligatures, Digraphs, and Alternate 798 Character Forms 800 [[anchor15: There is some internal redundancy and repetition in the 801 material in this section. Specific suggestions about to reduce or 802 eliminate redundant text would be appreciated. If no such 803 suggestions are received before -07 is posted, this note will be 804 removed.]] 806 Users often have expectations about character matching or equivalence 807 that are based on their own languages and the orthography of those 808 languages. These expectations may not be consistent with forms or 809 actions that can be naturally accommodated in a character coding 810 system, especially if multiple languages are written using the same 811 script but using different conventions. A Norwegian user might 812 expect a label with the ae-ligature to be treated as the same label 813 as one using the Swedish spelling with a-diaeresis even though 814 applying that mapping to English would be astonishing to users. A 815 user in German might expect a label with an o-umlaut and a label that 816 had "oe" substituted, but was otherwise the same, treated as 817 equivalent even though that substitution would be a clear error in 818 Swedish. A Chinese user might expect automatic matching of 819 Simplified and Traditional Chinese characters, but applying that 820 matching for Korean or Japanese text would create considerable 821 confusion. For that matter, an English user might expect "theater" 822 and "theatre" to match. 824 Related issues arise because there are a number of languages written 825 with alphabetic scripts in which single phonemes are written using 826 two characters, termed a "digraph", for example, the "ph" in 827 "pharmacy" and "telephone". (Note that characters paired in this 828 manner can also appear consecutively without forming a digraph, as in 829 "tophat".) Certain digraphs are normally indicated typographically 830 by setting the two characters closer together than they would be if 831 used consecutively to represent different phonemes. Some digraphs 832 are fully joined as ligatures (strictly designating setting totally 833 without intervening white space, although the term is sometimes 834 applied to close set pairs). An example of this may be seen when the 835 word "encyclopaedia" is set with a U+00E6 LATIN SMALL LIGATURE AE 836 (and some would not consider that word correctly spelled unless the 837 ligature form was used or the "a" was dropped entirely). When these 838 ligature and digraph forms have the same interpretation across all 839 languages that use a given script, application of Unicode 840 normalization generally resolves the differences and causes them to 841 match. When they have different interpretations, any requirements 842 for matching must utilize other methods, presumably at the registry 843 level, or users must be educated to understand that matching will not 844 occur. 846 Difficulties arise from the fact that a given ligature may be a 847 completely optional typographic convenience for representing a 848 digraph in one language (as in the above example with some spelling 849 conventions), while in another language it is a single character that 850 may not always be correctly representable by a two-letter sequence 851 (as in the above example with different spelling conventions). This 852 can be illustrated by many words in the Norwegian language, where the 853 "ae" ligature is the 27th letter of a 29-letter extended Latin 854 alphabet. It is equivalent to the 28th letter of the Swedish 855 alphabet (also containing 29 letters), U+00E4 LATIN SMALL LETTER A 856 WITH DIAERESIS, for which an "ae" cannot be substituted according to 857 current orthographic standards. 859 That character (U+00E4) is also part of the German alphabet where, 860 unlike in the Nordic languages, the two-character sequence "ae" is 861 usually treated as a fully acceptable alternate orthography for the 862 "umlauted a" character. The inverse is however not true, and those 863 two characters cannot necessarily be combined into an "umlauted a". 864 This also applies to another German character, the "umlauted o" 865 (U+00F6 LATIN SMALL LETTER O WITH DIAERESIS) which, for example, 866 cannot be used for writing the name of the author "Goethe". It is 867 also a letter in the Swedish alphabet where, like the "a with 868 diaeresis", it cannot be correctly represented as "oe" and in the 869 Norwegian alphabet, where it is represented, not as "o with 870 diaeresis", but as "slashed o", U+00F8. 872 Some of the ligatures that have explicit code points in Unicode were 873 given special handling in IDNA2003 and now pose additional problems 874 in transition. See Section 7.2. 876 Additional cases with alphabets written right to left are described 877 in Section 4.5. 879 Whether ligatures and digraphs are to be treated as a sequence of 880 characters or as a single standalone one constitute a problem that 881 cannot be resolved solely by operating on scripts. They are, 882 however, a key concern in the IDN context. Their satisfactory 883 resolution will require support in policies set by registries, which 884 therefore need to be particularly mindful not just of this specific 885 issue, but of all other related matters that cannot be dealt with on 886 an exclusively algorithmic and global basis. 888 Just as with the examples of different-looking characters that may be 889 assumed to be the same, it is in general impossible to deal with 890 these situations in a system such as IDNA -- or with Unicode 891 normalization generally -- since determining what to do requires 892 information about the language being used, context, or both. 893 Consequently, these specifications make no attempt to treat these 894 combined characters in any special way. However, their existence 895 provides a prime example of a situation in which a registry that is 896 aware of the language context in which labels are to be registered, 897 and where that language sometimes (or always) treats the two- 898 character sequences as equivalent to the combined form, should give 899 serious consideration to applying a "variant" model [RFC3743] 900 [RFC4290], or to prohibiting registration of one the forms entirely, 901 to reduce the opportunities for user confusion and fraud that would 902 result from the related strings being registered to different 903 parties. 905 [[anchor16: Placeholder: A discussion of the Arabic digit issue 906 should go here once it is resolved in some appropriate way.]] 908 4.4. Case Mapping and Related Issues 910 In the DNS, ASCII letters are stored with their case preserved. 911 Matching during the query process is case-independent, but none of 912 the information that might be represented by choices of case has been 913 lost. That model has been accidentally helpful because, as people 914 have created DNS labels by catenating words (or parts of words) to 915 form labels, case has often been used to distinguish among components 916 and make the labels more memorable. 918 The solution of keeping the characters separate but doing matching 919 independent of case is not feasible with IDNA or any IDNA-like model 920 because the matching would then have to be done on the server rather 921 than have characters mapped on the client. That situation was 922 recognized in IDNA2003 and nothing in these specifications 923 fundamentally changes it or could do so. In IDNA2003, all characters 924 are case-folded and mapped. That results in upper-case characters 925 being mapped to lower-case ones and in some other transformations of 926 alternate forms of characters, especially those that do not have (or 927 did not have) upper-case forms. For example, Greek Final Form Sigma 928 (U+03C2) is mapped to the medial form (U+03C3) and Eszett (German 929 Sharp S, U+00DF) is mapped to "ss". Neither of these mappings is 930 reversible because the upper case of U+03C3 is the Upper Case Sigma 931 (U+03A3) and "ss" is an ASCII string. IDNA2008 permits, at the risk 932 of some incompatibility, slightly more flexibility in this area by 933 avoid case folding and treating these characters as themselves. 934 Approaches to handling that incompatibility are discussed in 935 Section 7.2. Although information is lost in IDNA2003's ToASCII 936 operation so that, in some sense, neither Final Sigma nor Eszett can 937 be represented in an IDN at all, its guarantee of mapping when those 938 characters are used as input can be interpreted as violating one of 939 the conditions discussed in Section 7.4.1 and hence requiring a 940 prefix change. The consensus was to not make a prefix change in 941 spite of this issue. Of course, had a prefix change been made (at 942 the costs discussed in Section 7.4.3) there would have been several 943 options, including, if desired, assignment of the character to the 944 CONTEXTUAL RULE REQUIRED category and requiring that it only be used 945 in carefully-selected contexts. 947 4.5. Right to Left Text 949 In order to be sure that the directionality of right to left text is 950 unambiguous, IDNA2003 required that any label in which right to left 951 characters appear both starts and ends with them, not include any 952 characters with strong left to right properties (which excludes other 953 alphabetic characters but permits European digits), and rejects any 954 other string that contains a right to left character. This is one of 955 the few places where the IDNA algorithms (both in IDNA2003 and in 956 IDAN2008) are required to examine an entire label, not just 957 individual characters. The algorithmic model used in IDNA2003 958 rejects the label when the final character in a right to left string 959 requires a combining mark in order to be correctly represented. 961 That prohibition is not acceptable for writing systems for languages 962 written with consonantal alphabets to which diacritical vocalic 963 systems are applied, and for languages with orthographies derived 964 from them where the combining marks may have different functionality. 965 In both cases the combining marks can be essential components of the 966 orthography. Examples of this are Yiddish, written with an extended 967 Hebrew script, and Dhivehi (the official language of Maldives) which 968 is written in the Thaana script (which is, in turn, derived from the 969 Arabic script). IDNA2008 removes the restriction on final combining 970 characters with a new set of rules for right to left scripts and 971 their characters. Those new rules are specified in [IDNA2008-Bidi]. 973 5. IDNs and the Robustness Principle 975 The model of IDNs described in this document can be seen as a 976 particular instance of the "Robustness Principle" that has been so 977 important to other aspects of Internet protocol design. This 978 principle is often stated as "Be conservative about what you send and 979 liberal in what you accept" (See, e.g., Section 1.2.2 of the 980 applications-layer Host Requirements specification [RFC1123]). For 981 IDNs to work well, not only must the protocol be carefully designed 982 and implemented, but zone administrators (registries) must have and 983 require sensible policies about what is registered -- conservative 984 policies -- and implement and enforce them. 986 Conversely, lookup applications are expected to reject labels that 987 clearly violate global (protocol) rules (no one has ever seriously 988 claimed that being liberal in what is accepted requires being 989 stupid). However, once one gets past such global rules and deals 990 with anything sensitive to script or locale, it is necessary to 991 assume that garbage has not been placed into the DNS, i.e., one must 992 be liberal about what one is willing to look up in the DNS rather 993 than guessing about whether it should have been permitted to be 994 registered. 996 As mentioned elsewhere, if a string cannot be successfully found in 997 the DNS after the lookup processing described here, it makes no 998 difference whether it simply wasn't registered or was prohibited by 999 some rule at the registry. Applications should, however, be 1000 sensitive to the fact that, because of the possibility of DNS 1001 wildcards, the ability to successfully resolve a name does not 1002 guarantee that it was actually registered. 1004 If lookup applications, as a user interface (UI) or other local 1005 matter, decide to warn about some strings that are valid under the 1006 global rules but that they perceive as dangerous, that is their 1007 prerogative and we can only hope that the market (and maybe 1008 regulators) will reinforce the good choices and discourage the poor 1009 ones. In this context, a lookup application that decides a string 1010 that is valid under the protocol is dangerous and refuses to look it 1011 up is in violation of the protocols; one that is willing to look 1012 something up, but warns against it, is exercising a local choice. 1014 6. Front-end and User Interface Processing for Lookup 1016 Domain names may be identified and processed in many contexts. They 1017 may be typed in by users either by themselves or embedded in an 1018 identifier structured for a particular protocol or class of protocols 1019 such a email addresses, URIs, or IRIs. They may occur in running 1020 text or be processed by one system after being provided in another. 1021 Systems may wish to try to normalize URLs so as to determine (or 1022 guess) whether a reference is valid or two references point to the 1023 same object without actually looking the objects up and comparing 1024 them (that is necessary, not just a choice, for URI types that are 1025 not intended to be resolved). Some of these goals may be more easily 1026 and reliably satisfied than others. While there are strong arguments 1027 for any domain name that is placed "on the wire" -- transmitted 1028 between systems -- to be in the zero-ambiguity forms of A-labels, it 1029 is inevitable that programs that process domain names will encounter 1030 U-labels or variant forms. 1032 One source of such forms will be labels created under IDNA2003 1033 because that protocol allowed labels that were transformed from 1034 native-character format by mapping some characters into others before 1035 conversion into ACE ("xn--...") format. One consequence of the 1036 transformations was that, when the ToUnicode and ToASCII operations 1037 of IDNA2003 were applied, ToUnicode(ToASCII(original-label)) often 1038 did not produce the original label. IDNA2008 explicitly defines 1039 A-labels and U-labels as different forms of the same abstract label, 1040 forms that are stable when conversions are performed between them 1041 (without mappings). A different way of explaining this is that there 1042 are, today, domain names in files on the Internet that use characters 1043 that cannot be represented directly in, or recovered from, (A-label) 1044 domain names but for which interpretations are provided by IDNA2003. 1046 There are two major categories of such characters, those that are 1047 removed by NFKC normalization and those upper-case characters that 1048 are mapped to lower-case (there are also a few characters that are 1049 given special-case mapping treatment in Stringprep, including lower- 1050 case characters that are case-folded into other lower-case characters 1051 or strings). 1053 Other issues in domain name identification and processing arise 1054 because IDNA2003 specified that several other characters be treated 1055 as equivalent to the ASCII period (dot, full stop) character used as 1056 a label separator. If a string that might be a domain name appears 1057 in an arbitrary context (such as running text), it is difficult, even 1058 with only ASCII characters, to know whether an actual domain name (or 1059 a protocol parameter like a URI) is present and where it starts and 1060 ends. When using Unicode, this gets even more difficult if treatment 1061 of certain special characters (like the dot that separates labels in 1062 a domain name) depends on context (e.g., prior knowledge of whether 1063 the string represents a domain name or not). That knowledge is not 1064 available if the primary heuristic for identifying the presence of 1065 domain names in strings depends on the presence of dots separating 1066 groups of characters with no intervening spaces. 1068 As discussed elsewhere in this document, the IDNA2008 model removes 1069 all of these mappings and interpretations, including the equivalence 1070 of different forms of dots, from the protocol, discouraging such 1071 mappings and leaving them, when necessary, to local processing. This 1072 should not be taken to imply that local processing is optional or can 1073 be avoided entirely, even if doing so might have been desirable in a 1074 world without IDNA2003 IDNs in files and archives. Instead, unless 1075 the program context is such that it is known that any IDNs that 1076 appear will contain either U-label or A-label forms, or that other 1077 forms can safely be rejected, some local processing of apparent 1078 domain name strings will be required, both to maintain compatibility 1079 with IDNA2003 and to prevent user astonishment. Such local 1080 processing, while not specified in this document or the associated 1081 ones, will generally take one of two forms: 1083 o Generic Preprocessing. 1084 When the context in which the program or system that processes 1085 domain names operates is global, a reasonable balance must be 1086 found that is sensitive to the broad range of local needs and 1087 assumptions while, at the same time, not sacrificing the needs of 1088 one language, script, or user population to those of another. 1090 For this case, the best practice will usually be to apply NFKC and 1091 case-mapping (or, perhaps better yet, Stringprep itself), plus 1092 dot-mapping where appropriate, to the domain name string prior to 1093 applying IDNA. That practice will not only yield a reasonable 1094 compromise of user experience with protocol requirements but will 1095 be almost completely compatible with the various forms permitted 1096 by IDNA2003. 1098 o Highly Localized Preprocessing. 1099 Unlike the case above, there will be some situations in which 1100 software will be highly localized for a particular environment and 1101 carefully adapted to the expectations of users in that 1102 environment. The many discussions about using the Internet to 1103 preserve and support local cultures suggest that these cases may 1104 be more common in the future than they have been so far. 1106 In these cases, we should avoid trying to tell implementers what 1107 they should accept, if only because they are quite likely (and for 1108 good reason) to ignore us. We would assume that they would map 1109 characters that the intuitions of their users would suggest be 1110 mapped and would hope that they would do that mapping as early as 1111 possible, storing A-label or U-label forms in files and 1112 transporting only those forms between systems. One can imagine 1113 switches about whether some sorts of mappings occur, warnings 1114 before applying them or, in a slightly more extreme version of the 1115 approach taken in Internet Explorer version 7 (IE7), systems that 1116 utterly refuse to handle "strange" characters at all if they 1117 appear in U-label form. None of those local decisions are a 1118 threat to interoperability as long as (i) only U-labels and 1119 A-labels are used in interchange with systems outside the local 1120 environment, (ii) no character that would be valid in a U-label as 1121 itself is mapped to something else, (iii) any local mappings are 1122 applied as a preprocessing step (or, for conversions from U-labels 1123 or A-labels to presentation forms, postprocessing), not as part of 1124 IDNA processing proper, and (iv) appropriate consideration is 1125 given to labels that might have entered the environment in 1126 conformance to IDNA2003. 1128 In either case, it is vital that user interface designs and, where 1129 the interfaces are not sufficient, users, be aware that the only 1130 forms of domain names that this protocol anticipates will resolve 1131 globally or compare equal when crude methods (i.e., those not 1132 conforming to the strict definition of label equivalence given in 1133 [IDNA2008-Defs]) are used are those in which all native-script labels 1134 are in U-label form. Forms that assume mapping will occur, 1135 especially forms that were not valid under IDNA2003, may or may not 1136 function in predictable ways across all implementations. 1138 User interfaces involving Latin-based scripts should take special 1139 care when considering how to handle case mapping because small 1140 differences in label strings may cause behavior that is astonishing 1141 to users. Because case-insensitive comparison is done for ASCII 1142 strings by DNS-servers, an all-ASCII label is treated as case- 1143 insensitive. However, if even one of the characters of that string 1144 is replaced by one that requires the label to be given IDN treatment 1145 (e.g., by adding a diacritical mark), then the label effectively 1146 becomes case-sensitive because only lower-case characters are 1147 permitted in IDNs. This suggests that case mapping for Latin-based 1148 scripts (and possibly other scripts with case distinctions) as a 1149 preprocessing matter in applications may be wise to prevent user 1150 astonishment, but, since all applications may not do this and 1151 ambiguity in transport is not desirable, the that case-dependent 1152 forms should not be stored in files. 1154 The comments above apply only in operations that look up names or 1155 interpret files. There are several reasons why registration 1156 activities should require final names and verification of those names 1157 by the would-be registrant. 1159 7. Migration from IDNA2003 and Unicode Version Synchronization 1161 7.1. Design Criteria 1163 As mentioned above and in RFC 4690, two key goals of the IDNA2008 1164 design are to enable applications to be agnostic about whether they 1165 are being run in environments supporting any Unicode version from 3.2 1166 onward and to permit incrementally adding new characters, character 1167 groups, scripts, and other character collections as they are 1168 incorporated into Unicode, without disruption and, in the long term, 1169 without "heavy" processes such as those involving IETF consensus. 1170 (An IETF consensus process is required by the IDNA2008 specifications 1171 and is expected to be required and used until significant experience 1172 accumulates with IDNA operations and new versions of Unicode.) The 1173 mechanisms that support this are outlined above and elsewhere in the 1174 IDNA2008 document set, but this section reviews them in a context 1175 that may be more helpful to those who need to understand the approach 1176 and make plans for it. 1178 7.1.1. General IDNA Validity Criteria 1180 The general criteria for a putative label, and the collection of 1181 characters that make it up, to be considered IDNA-valid are (the 1182 actual rules are rigorously defined in the "Protocol" and "Tables" 1183 documents): 1185 o The characters are "letters", marks needed to form letters, 1186 numerals, or other code points used to write words in some 1187 language. Symbols, drawing characters, and various notational 1188 characters are intended to be permanently excluded -- some because 1189 they are harmful in URI, IRI, or similar contexts (e.g., 1190 characters that appear to be slashes or other reserved URI 1191 punctuation) and others because there is no evidence that they are 1192 important enough to Internet operations or internationalization to 1193 justify expansion of domain names beyond the general principle of 1194 "letters, digits, and hyphen" and the complexities that would come 1195 with it (additional discussion and rationale for the symbol 1196 decision appears in Section 7.6). 1198 o Other than in very exceptional cases, e.g., where they are needed 1199 to write substantially any word of a given language, punctuation 1200 characters are excluded as well. The fact that a word exists is 1201 not proof that it should be usable in a DNS label and DNS labels 1202 are not expected to be usable for multiple-word phrases (although 1203 they are certainly not prohibited if the conventions and 1204 orthography of a particular language cause that to be possible). 1205 Even for English, very common constructions -- contractions like 1206 "don't" or "it's", names that are written with apostrophes such as 1207 "O'Reilly", or characters for which apostrophes are common 1208 substitutes cannot be represented in DNS labels. Words in English 1209 whose usually-preferred spellings include diacritical marks cannot 1210 be represented under the original hostname rules, but most can be 1211 represented if treated as IDNs. 1213 o Characters that are unassigned (have no character assignment at 1214 all) in the version of Unicode being used by the registry or 1215 application are not permitted, even on lookup. The issues 1216 involved in this decision are discussed in Section 7.7. 1218 o Any character that is mapped to another character by a current 1219 version of NFKC is prohibited as input to IDNA (for either 1220 registration or lookup). With a few exceptions, this principle 1221 excludes any character mapped to another by Nameprep [RFC3491]. 1223 Tables used to identify the characters that are IDNA-valid are 1224 expected to be driven by the principles above, principles that are 1225 specified exactly in [IDNA2008-Tables]). The rules given there are 1226 normative, rather than being just an interpretation of the tables. 1228 7.1.2. Labels in Registration 1230 Anyone entering a label into a DNS zone must properly validate that 1231 label -- i.e., be sure that the criteria for that label are met -- in 1232 order for applications to work as intended. This principle is not 1233 new. For example, since the DNS was first deployed, zone 1234 administrators have been expected to verify that names meet 1235 "hostname" [RFC0952] where necessary for the expected applications. 1236 Later addition of special service location formats [RFC2782] imposed 1237 new requirements on zone administrators for the use of labels that 1238 conform to the requirements of those formats. For zones that will 1239 contain IDNs, support for Unicode version-independence requires 1240 restrictions on all strings placed in the zone. In particular, for 1241 such zones: 1243 o Any label that appears to be an A-label, i.e., any label that 1244 starts in "xn--", must be IDNA-valid, i.e., they must be valid 1245 A-labels, as discussed in Section 2 above. 1247 o The Unicode tables (i.e., tables of code points, character 1248 classes, and properties) and IDNA tables (i.e., tables of 1249 contextual rules such as those that appear in the Tables 1250 document), must be consistent on the systems performing or 1251 validating labels to be registered. Note that this does not 1252 require that tables reflect the latest version of Unicode, only 1253 that all tables used on a given system are consistent with each 1254 other. 1256 Under this model, a registry (or entity communicating with a registry 1257 to accomplish name registrations) will need to update its tables -- 1258 both the Unicode-associated tables and the tables of permitted IDN 1259 characters -- to enable a new script or other set of new characters. 1260 It will not be affected by newer versions of Unicode, or newly- 1261 authorized characters, until and unless it wishes to make those 1262 registrations. The zone administrator is also responsible -- under 1263 the protocol and to registrants and users -- for both checking as 1264 required by the protocol and verification that whatever policies it 1265 develops are complied with, whether those policies are for minimizing 1266 risks due to confusable characters and sequences, for preserving 1267 language or script integrity, or for other purposes. Those checking 1268 and verification procedures are more extensive than those that are is 1269 expected of applications systems that look names up. 1271 Systems looking up or resolving DNS labels, especially IDN DNS 1272 labels, must be able to assume that applicable registration rules 1273 were followed for names entered into the DNS. 1275 7.1.3. Labels in Lookup 1277 Anyone looking up a label in a DNS zone is required to 1279 o Maintain a consistent set of tables, as discussed above. As with 1280 registration, the tables need not reflect the latest version of 1281 Unicode but they must be consistent. 1283 o Validate the characters in labels to be looked up only to the 1284 extent of determining that the U-label does not contain either 1285 code points prohibited by IDNA (categorized as "DISALLOWED") or 1286 code points that are unassigned in its version of Unicode. 1288 o Validate the label itself for conformance with a small number of 1289 whole-label rules, notably verifying that there are no leading 1290 combining marks, that the "bidi" conditions are met if right to 1291 left characters appear, that any required contextual rules are 1292 available and that, if such rules are associated with Joiner 1293 Controls, they are tested. 1295 o Avoid validating other contextual rules about characters, 1296 including mixed-script label prohibitions, although such rules may 1297 be used to influence presentation decisions in the user interface. 1298 [[anchor20: Check this, and all similar statements, against 1299 Protocol when that is finished.]] 1301 By avoiding applying its own interpretation of which labels are valid 1302 as a means of rejecting lookup attempts, the lookup application 1303 becomes less sensitive to version incompatibilities with the 1304 particular zone registry associated with the domain name. 1306 An application or client that processes names according to this 1307 protocol and then resolves them in the DNS will be able to locate any 1308 name that is validly registered, as long as its version of the 1309 Unicode-associated tables is sufficiently up-to-date to interpret all 1310 of the characters in the label. Messages to users should distinguish 1311 between "label contains an unallocated code point" and other types of 1312 lookup failures. A failure on the basis of an old version of Unicode 1313 may lead the user to a desire to upgrade to a newer version, but will 1314 have no other ill effects (this is consistent with behavior in the 1315 transition to the DNS when some hosts could not yet handle some forms 1316 of names or record types). 1318 7.2. Changes in Character Interpretations 1320 [[anchor21: Note in Draft: This subsection is completely new in 1321 version -04 and has been further tuned in -05 and -06 of this 1322 document. It could almost certainly use improvement, although this 1323 note will be removed if there are not significant suggestions about 1324 the -06 version. It also contains some material that is redundant 1325 with material in other sections. I have not tried to remove that 1326 material and will not do so until the WG concludes that this section 1327 is relatively stable, but would appreciate help in identifying what 1328 should be removed or how this might be enhanced to contain more of 1329 that other material. --JcK]] 1331 In those scripts that make case distinctions, there are a few 1332 characters for which an obvious and unique upper case character has 1333 not historically been available to match a lower case one or vice 1334 versa. For those characters, the mappings used in constructing the 1335 Stringprep tables for IDNA2003, performed using the Unicode CaseFold 1336 operation (See Section 5.8 of the Unicode Standard [Unicode51]), 1337 generate different characters or sets of characters. Those 1338 operations are not reversible and lose even more information than 1339 traditional upper case or lower case transformations, but are more 1340 useful than those transformations for comparison purposes. Two 1341 notable characters of this type are the German character Eszett 1342 (Sharp S, U+00DF) and the Greek Final Form Sigma (U+03C2). The 1343 former is case-folded to the ASCII string "ss", the latter to a 1344 medial (Lower Case) Sigma (U+03C3). 1346 The decision to eliminate mappings, including case folding, from the 1347 IDNA2008 protocol in order to make A-labels and U-labels idempotent 1348 made these characters problematic. If they were to be disallowed, 1349 important words and mnemonics could not be written in 1350 orthographically reasonable ways. If they were to be permitted as 1351 characters distinct from the forms produced by case folding, there 1352 would be no information loss and registries would have maximum 1353 flexibility, but labels using those characters that were looked up 1354 according to IDNA2003 rules would be transformed into A-labels using 1355 their case-mapped variations while lookup according to IDNA2008 rules 1356 would be based on different A-labels that represented the actual 1357 characters. 1359 With the understanding that there would be incompatibility either way 1360 but a judgment that the incompatibility was not significant enough to 1361 just a prefix change, the WG concluded that Eszett and Final Form 1362 Sigma should be treated as distinct and Protocol-Valid characters. 1364 The decision faces registries, especially registries maintaining 1365 zones for third parties, with a variation on what has become a 1366 familiar problem: how to introduce a new service in a way that does 1367 not create confusion or significantly weaken or invalidate existing 1368 identifiers. 1370 There have traditionally been several approaches to problems of this 1371 type. Without any preference or claim to completeness, these are: 1373 o Do not permit use of the newly-available character at the registry 1374 level. This might cause lookup failures if a domain name were to 1375 be written with the expectation of the IDNA2003 mapping behavior, 1376 but would eliminate any possibility of false matches. 1378 o Hold a "sunrise"-like arrangement in which holders of the 1379 previously-mapped labels (labels containing "ss" in the Eszett 1380 case or ones containing Lower Case Sigma in the Final Sigma case) 1381 are given priority (and perhaps other benefits) for registering 1382 the corresponding string containing the newly-available 1383 characters. 1385 o Adopt some sort of "variant" approach in which registrants either 1386 obtained labels with both character forms or one of them was 1387 blocked from registration by anyone but the registrant of the 1388 other form. 1390 In principle, lookup applications could also compensate for the 1391 difference in interpretation by looking up the string according to 1392 the interpretation specified in these documents and then, if that 1393 failed, doing the lookup with the mapping, simulating the IDNA2003 1394 interpretation. The risk of false positives is such that this is 1395 generally to be discouraged unless the application is able to engage 1396 in a "is this what you meant" dialogue with the end user. 1398 7.3. More Flexibility in User Agents 1400 These documents do not specify mappings between one character or code 1401 point and others for any reason. Instead, they prohibit the 1402 characters that would be mapped to others by normalization, upper 1403 case to lower case changes, or other rules. As examples, while 1404 mathematical characters based on Latin ones are accepted as input to 1405 IDNA2003, they are prohibited in IDNA2008. Similarly, double-width 1406 characters and other variations are prohibited as IDNA input. 1408 Since the rules in [IDNA2008-Tables] have the effect that only 1409 strings that are not transformed by NFKC are valid, if an application 1410 chooses to perform NFKC normalization before lookup, that operation 1411 is safe since this will never make the application unable to look up 1412 any valid string. However, as discussed above, the application 1413 cannot guarantee that any other application will perform that 1414 mapping, so it should be used only with caution and for informed 1415 users. 1417 In many cases these prohibitions should have no effect on what the 1418 user can type as input to the lookup process. It is perfectly 1419 reasonable for systems that support user interfaces to perform some 1420 character mapping that is appropriate to the local environment. This 1421 would normally be done prior to actual invocation of IDNA. At least 1422 conceptually, the mapping would be part of the Unicode conversions 1423 discussed above and in [IDNA2008-Protocol]. However, those changes 1424 will be local ones only -- local to environments in which users will 1425 clearly understand that the character forms are equivalent. For use 1426 in interchange among systems, it appears to be much more important 1427 that U-labels and A-labels can be mapped back and forth without loss 1428 of information. 1430 One specific, and very important, instance of this strategy arises 1431 with case-folding. In the ASCII-only DNS, names are looked up and 1432 matched in a case-independent way, but no actual case-folding occurs. 1433 Names can be placed in the DNS in either upper or lower case form (or 1434 any mixture of them) and that form is preserved, returned in queries, 1435 and so on. IDNA2003 simulated that behavior for non-ASCII strings by 1436 performing case-folding at registration time (resulting in only 1437 lower-case IDNs in the DNS) and when names were looked up. 1439 As suggested earlier in this section, it appears to be desirable to 1440 do as little character mapping as possible consistent with having 1441 Unicode work correctly (e.g., NFC mapping to resolve different 1442 codings for the same character is still necessary although the 1443 specifications require that it be performed prior to invoking the 1444 protocol) and to make the mapping between A-labels and U-labels 1445 idempotent. Case-mapping is not an exception to this principle. If 1446 only lower case characters can be registered in the DNS (i.e., be 1447 present in a U-label), then IDNA2008 should prohibit upper-case 1448 characters as input (and therefore does so). Some other 1449 considerations reinforce this conclusion. For example, an essential 1450 element of the ASCII case-mapping functions is that, for individual 1451 characters, uppercase(character) must be equal to 1452 uppercase(lowercase(character)). That requirement may not be 1453 satisfied with IDNs. For example, there are some characters in 1454 scripts that use case distinction that do not have counterparts in 1455 one case or the other. The relationship between upper case and lower 1456 case may even be language-dependent, with different languages (or 1457 even the same language in different areas) expecting different 1458 mappings. Of course, the expectations of users who are accustomed to 1459 a case-insensitive DNS environment will probably be well-served if 1460 user agents perform case folding prior to IDNA processing, but the 1461 IDNA procedures themselves should neither require such mapping nor 1462 expect them when they are not natural to the localized environment. 1464 7.4. The Question of Prefix Changes 1466 The conditions that would require a change in the IDNA ACE prefix 1467 ("xn--" for the version of IDNA specified in [RFC3490]) have been a 1468 great concern to the community. A prefix change would clearly be 1469 necessary if the algorithms were modified in a manner that would 1470 create serious ambiguities during subsequent transition in 1471 registrations. This section summarizes our conclusions about the 1472 conditions under which changes in prefix would be necessary and the 1473 implications of such a change. 1475 7.4.1. Conditions Requiring a Prefix Change 1477 An IDN prefix change is needed if a given string would be looked up 1478 or otherwise interpreted differently depending on the version of the 1479 protocol or tables being used. Consequently, work to update IDNs 1480 would require a prefix change if, and only if, one of the following 1481 four conditions were met: 1483 1. The conversion of an A-label to Unicode (i.e., a U-label) yields 1484 one string under IDNA2003 (RFC3490) and a different string under 1485 IDNA2008. 1487 2. An input string that is valid under IDNA2003 and also valid under 1488 IDNA2008 yields two different A-labels with the different 1489 versions of IDNA. This condition is believed to be essentially 1490 equivalent to the one above except for a very small number of 1491 edge cases which may not, pragmatically, justify a prefix change 1492 (See Section 7.2). 1494 Note, however, that if the input string is valid under one 1495 version and not valid under the other, this condition does not 1496 apply. See the first item in Section 7.4.2, below. 1498 3. A fundamental change is made to the semantics of the string that 1499 is inserted in the DNS, e.g., if a decision were made to try to 1500 include language or specific script information in that string, 1501 rather than having it be just a string of characters. 1503 4. A sufficiently large number of characters is added to Unicode so 1504 that the Punycode mechanism for block offsets no longer has 1505 enough capacity to reference the higher-numbered planes and 1506 blocks. This condition is unlikely even in the long term and 1507 certain not to arise in the next few years. 1509 7.4.2. Conditions Not Requiring a Prefix Change 1511 In particular, as a result of the principles described above, none of 1512 the following changes require a new prefix: 1514 1. Prohibition of some characters as input to IDNA. This may make 1515 names that are now registered inaccessible, but does not require 1516 a prefix change. 1518 2. Adjustments in IDNA tables or actions, including normalization 1519 definitions, that affect characters that were already invalid 1520 under IDNA2003. 1522 3. Changes in the style of the IDNA definition that does not alter 1523 the actions performed by IDNA. 1525 7.4.3. Implications of Prefix Changes 1527 While it might be possible to make a prefix change, the costs of such 1528 a change are considerable. Even if they wanted to do so, registries 1529 could not convert all IDNA2003 ("xn--") registrations to a new form 1530 at the same time and synchronize that change with applications 1531 supporting lookup. Unless all existing registrations were simply to 1532 be declared invalid (and perhaps even then) systems that needed to 1533 support both labels with old prefixes and labels with new ones would 1534 first process a putative label under the IDNA2008 rules and try to 1535 look it up and then, if it were not found, would process the label 1536 under IDNA2003 rules and look it up again. That process could 1537 significantly slow down all processing that involved IDNs in the DNS 1538 especially since, in principle, a fully-qualified name could contain 1539 a mixture of labels that were registered with the old and new 1540 prefixes, a situation that would make the use of DNS caching very 1541 difficult. In addition, looking up the same input string as two 1542 separate A-labels would create some potential for confusion and 1543 attacks, since they could, in principle, map to different targets and 1544 then resolve to different entries in the DNS. 1546 Consequently, a prefix change is to be avoided if at all possible, 1547 even if it means accepting some IDNA2003 decisions about character 1548 distinctions as irreversible and/or giving special treatment to edge 1549 cases. 1551 7.5. Stringprep Changes and Compatibility 1553 The Nameprep [RFC3491] specification, a key part of IDNA2003, is a 1554 profile of Stringprep [RFC3454]. While Nameprep is a Stringprep 1555 profile specific to IDNA, Stringprep is used by a number of other 1556 protocols. Concerns have been expressed about problems for non-DNS 1557 uses of Stringprep being caused by changes to the specification 1558 intended to improve the handling of IDNs, most notably as this might 1559 affect identification and authentication protocols. The proposed new 1560 inclusion tables [IDNA2008-Tables], the reduction in the number of 1561 characters permitted as input for registration or lookup (Section 3), 1562 and even the proposed changes in handling of right to left strings 1563 [IDNA2008-Bidi] either give interpretations to strings prohibited 1564 under IDNA2003 or prohibit strings that IDNA2003 permitted. The 1565 IDNA2008 protocol does not use either Nameprep or Stringprep at all, 1566 so there are no side-effect changes to other protocols. 1568 It is particularly important to keep IDNA processing separate from 1569 processing for various security protocols because some of the 1570 constraints that are necessary for smooth and comprehensible use of 1571 IDNs may be unwanted or undesirable in other contexts. For example, 1572 the criteria for good passwords or passphrases are very different 1573 from those for desirable IDNs: passwords should be hard to guess, 1574 while domain names should normally be easily memorable. Similarly, 1575 internationalized SCSI identifiers and other protocol components are 1576 likely to have different requirements than IDNs. 1578 7.6. The Symbol Question 1580 One of the major differences between this specification and the 1581 original version of IDNA is that the original version permitted non- 1582 letter symbols of various sorts, including punctuation and line- 1583 drawing symbols, in the protocol. They were always discouraged in 1584 practice. In particular, both the "IESG Statement" about IDNA and 1585 all versions of the ICANN Guidelines specify that only language 1586 characters be used in labels. This specification disallows symbols 1587 entirely. There are several reasons for this, which include: 1589 o As discussed elsewhere, the original IDNA specification assumed 1590 that as many Unicode characters as possible should be permitted, 1591 directly or via mapping to other characters, in IDNs. This 1592 specification operates on an inclusion model, extrapolating from 1593 the LDH rules -- which have served the Internet very well -- to a 1594 Unicode base rather than an ASCII base. 1596 o Most Unicode names for letters are, in most cases, fairly 1597 intuitive, unambiguous and recognizable to users of the relevant 1598 script. Symbol names are more problematic because there may be no 1599 general agreement on whether a particular glyph matches a symbol; 1600 there are no uniform conventions for naming; variations such as 1601 outline, solid, and shaded forms may or may not exist; and so on. 1602 As just one example, consider a "heart" symbol as it might appear 1603 in a logo that might be read as "I love...". While the user might 1604 read such a logo as "I love..." or "I heart...", considerable 1605 knowledge of the coding distinctions made in Unicode is needed to 1606 know that there more than one "heart" character (e.g., U+2665, 1607 U+2661, and U+2765) and how to describe it. These issues are of 1608 particular importance if strings are expected to be understood or 1609 transcribed by the listener after being read out loud. 1610 [[anchor22: The above paragraph remains controversial as to 1611 whether it is valid. The WG will need to make a decision if this 1612 section is not dropped entirely.]] 1614 o Consider the case of a screen reader used by blind Internet users 1615 who must listen to renderings of IDN domain names and possibly 1616 reproduce them on the keyboard. 1618 o As a simplified example of this, assume one wanted to use a 1619 "heart" or "star" symbol in a label. This is problematic because 1620 those names are ambiguous in the Unicode system of naming (the 1621 actual Unicode names require far more qualification). A user or 1622 would-be registrant has no way to know -- absent careful study of 1623 the code tables -- whether it is ambiguous (e.g., where there are 1624 multiple "heart" characters) or not. Conversely, the user seeing 1625 the hypothetical label doesn't know whether to read it -- try to 1626 transmit it to a colleague by voice -- as "heart", as "love", as 1627 "black heart", or as any of the other examples below. 1629 o The actual situation is even worse than this. There is no 1630 possible way for a normal, casual, user to tell the difference 1631 between the hearts of U+2665 and U+2765 and the stars of U+2606 1632 and U+2729 or the without somehow knowing to look for a 1633 distinction. We have a white heart (U+2661) and few black hearts. 1634 Consequently, describing a label as containing a heart hopelessly 1635 ambiguous: we can only know that it contains one of several 1636 characters that look like hearts or have "heart" in their names. 1637 In cities where "Square" is a popular part of a location name, one 1638 might well want to use a square symbol in a label as well and 1639 there are far more squares of various flavors in Unicode than 1640 there are hearts or stars. 1642 o The consequence of these ambiguities of description and 1643 dependencies on distinctions that were, or were not, made in 1644 Unicode codings is that symbols are a very poor basis for reliable 1645 communication. Consistent with this conclusion, the Unicode 1646 standard recommends that strings used in identifiers not contain 1647 symbols or punctuation [Unicode-UAX31]. Of course, these 1648 difficulties with symbols do not arise with actual pictographic 1649 languages and scripts which would be treated like any other 1650 language characters; the two should not be confused. 1652 7.7. Migration Between Unicode Versions: Unassigned Code Points 1654 In IDNA2003, labels containing unassigned code points are looked up 1655 on the assumption that, if they appear in labels and can be mapped 1656 and then resolved, the relevant standards must have changed and the 1657 registry has properly allocated only assigned values. 1659 In the protocol as described in these documents, strings containing 1660 unassigned code points must not be either looked up or registered. 1661 There are several reasons for this, with the most important ones 1662 being: 1664 o It cannot be known in advance, and with sufficient reliability, 1665 that a code point that was not previously assigned will not be 1666 assigned to a compatibility character or one that would be 1667 otherwise disallowed by the rules in [IDNA2008-Tables]. In 1668 IDNA2003, since there is no direct dependency on NFKC 1669 (Stringprep's tables are based on NFKC, but IDNA2003 depends only 1670 on Stringprep), allocation of a compatibility character might 1671 produce some odd situations, but it would not be a problem. In 1672 IDNA2008, where compatibility characters are assigned to 1673 DISALLOWED unless character-specific exceptions are made, 1674 permitting strings containing unassigned characters to be looked 1675 up would permit violating the principle that characters in 1676 DISALLOWED are not looked up. 1678 o The Unicode Standard specifies that an unassigned code point 1679 normalizes (and, where relevant, case folds) to itself. If the 1680 code point is later assigned to a character, and particularly if 1681 the newly-assigned code point has a combining class that 1682 determines its placement relative to other combining characters, 1683 it could normalize to some other code point or sequence, creating 1684 confusion and/or violating other rules listed here. 1686 o Tests involving the context of characters (e.g., some characters 1687 being permitted only adjacent to ones of specific types but 1688 otherwise invisible or very problematic for other reasons) and 1689 integrity tests on complete labels are needed. Unassigned code 1690 points cannot be permitted because one cannot determine whether 1691 particular code points will require contextual rules (and what 1692 those rules should be) before characters are assigned to them and 1693 the properties of those characters fully understood. 1695 o More generally, the status of an unassigned character with regard 1696 to the DISALLOWED and PROTOCOL-VALID categories, and whether 1697 contextual rules are required with the latter, cannot be evaluated 1698 until a character is actually assigned and known. By contrast, 1699 characters that are actually DISALLOWED are placed in that 1700 category only as a consequence of rules applied to known 1701 properties or per-character evaluation. 1703 Another way to look at this is that permitting an unassigned 1704 character to be looked up is nearly equivalent to reclassifying a 1705 character from DISALLOWED to PROTOCOL-VALID since different systems 1706 will interpret the character in different ways. 1708 It is possible to argue that the issues above are not important and 1709 that, as a consequence, it is better to retain the principle of 1710 looking up labels even if they contain unassigned characters because 1711 all of the important scripts and characters have been coded as of 1712 Unicode 5.1 and hence unassigned code points will be assigned only to 1713 obscure characters or archaic scripts. Unfortunately, that does not 1714 appear to be a safe assumption for at least two reasons. First, much 1715 the same claim of completeness has been made for earlier versions of 1716 Unicode. The reality is that a script that is obscure to much of the 1717 world may still be very important to those who use it. Cultural and 1718 linguistic preservation principles make it inappropriate to declare 1719 the script of no importance in IDNs. Second, we already have 1720 counterexamples in, e.g., the relationships associated with new Han 1721 characters being added (whether in the BMP or in Unicode Plane 2). 1723 7.8. Other Compatibility Issues 1725 The 2003 IDNA model includes several odd artifacts of the context in 1726 which it was developed. Many, if not all, of these are potential 1727 avenues for exploits, especially if the registration process permits 1728 "source" names (names that have not been processed through IDNA and 1729 Nameprep) to be registered. As one example, since the character 1730 Eszett, used in German, is mapped by IDNA2003 into the sequence "ss" 1731 rather than being retained as itself or prohibited, a string 1732 containing that character but that is otherwise in ASCII is not 1733 really an IDN (in the U-label sense defined above) at all. After 1734 Nameprep maps the Eszett out, the result is an ASCII string and so 1735 does not get an xn-- prefix, but the string that can be displayed to 1736 a user appears to be an IDN. The newer version of the protocol 1737 eliminates this artifact. A character is either permitted as itself 1738 or it is prohibited; special cases that make sense only in a 1739 particular linguistic or cultural context can be dealt with as 1740 localization matters where appropriate. 1742 8. Name Server Considerations 1744 8.1. Processing Non-ASCII Strings 1746 Existing DNS servers do not know the IDNA rules for handling non- 1747 ASCII forms of IDNs, and therefore need to be shielded from them. 1748 All existing channels through which names can enter a DNS server 1749 database (for example, master files (as described in RFC 1034) and 1750 DNS update messages [RFC2136]) are IDN-unaware because they predate 1751 IDNA. Other sections of this document provide the needed shielding 1752 by ensuring that internationalized domain names entering DNS server 1753 databases through such channels have already been converted to their 1754 equivalent ASCII A-label forms. 1756 Because of the distinction made between the algorithms for 1757 Registration and Lookup in [IDNA2008-Protocol] (a domain name 1758 containing only ASCII codepoints can not be converted to an A-label), 1759 there can not be more than one A-label form for any given U-label. 1761 As specified in RFC 2181 [RFC2181], the DNS protocol explicitly 1762 allows domain labels to contain octets beyond the ASCII range 1763 (0000..007F), and this document does not change that. Note, however, 1764 that there is no defined interpretation of octets 0080..00FF as 1765 characters. If labels containing these octets are returned to 1766 applications, unpredictable behavior could result. The A-label form, 1767 which cannot contain those characters, is the only standard 1768 representation for internationalized labels in the DNS protocol. 1770 8.2. DNSSEC Authentication of IDN Domain Names 1772 DNS Security (DNSSEC) [RFC2535] is a method for supplying 1773 cryptographic verification information along with DNS messages. 1774 Public Key Cryptography is used in conjunction with digital 1775 signatures to provide a means for a requester of domain information 1776 to authenticate the source of the data. This ensures that it can be 1777 traced back to a trusted source, either directly or via a chain of 1778 trust linking the source of the information to the top of the DNS 1779 hierarchy. 1781 IDNA specifies that all internationalized domain names served by DNS 1782 servers that cannot be represented directly in ASCII MUST use the 1783 A-label form. Conversion to A-labels MUST be performed prior to a 1784 zone being signed by the private key for that zone. Because of this 1785 ordering, it is important to recognize that DNSSEC authenticates a 1786 domain name containing A-labels or conventional LDH-labels, not 1787 U-labels. In the presence of DNSSEC, no form of a zone file or query 1788 response that contains a U-label may be signed or the signature 1789 validated. 1791 One consequence of this for sites deploying IDNA in the presence of 1792 DNSSEC is that any special purpose proxies or forwarders used to 1793 transform user input into IDNs must be earlier in the lookup flow 1794 than DNSSEC authenticating nameservers for DNSSEC to work. 1796 8.3. Root and other DNS Server Considerations 1798 IDNs in A-label form will generally be somewhat longer than current 1799 domain names, so the bandwidth needed by the root servers is likely 1800 to go up by a small amount. Also, queries and responses for IDNs 1801 will probably be somewhat longer than typical queries historically, 1802 so EDNS0 [RFC2671] support may be more important (otherwise, queries 1803 and responses may be forced to go to TCP instead of UDP). 1805 9. Internationalization Considerations 1807 DNS labels and fully-qualified domain names provide mnemonics that 1808 assist in identifying and referring to resources on the Internet. 1809 IDNs expand the range of those mnemonics to include those based on 1810 languages and character sets other than Western European and Roman- 1811 derived ones. But domain "names" are not, in general, words in any 1812 language. The recommendations of the IETF policy on character sets 1813 and languages, BCP 18 [RFC2277] are applicable to situations in which 1814 language identification is used to provide language-specific 1815 contexts. The DNS is, by contrast, global and international and 1816 ultimately has nothing to do with languages. Adding languages (or 1817 similar context) to IDNs generally, or to DNS matching in particular, 1818 would imply context dependent matching in DNS, which would be a very 1819 significant change to the DNS protocol itself. It would also imply 1820 that users would need to identify the language associated with a 1821 particular label in order to look that label up, a decision that 1822 would be impossible in many or most cases. 1824 10. IANA Considerations 1826 This section gives an overview of registries required for IDNA. The 1827 actual definitions of the first two appear in [IDNA2008-Tables]. 1829 10.1. IDNA Character Registry 1831 The distinction among the three major categories "UNASSIGNED", 1832 "DISALLOWED", and "PROTOCOL-VALID" is made by special categories and 1833 rules that are integral elements of [IDNA2008-Tables]. Convenience 1834 in programming and validation requires a registry of characters and 1835 scripts and their categories, updated for each new version of Unicode 1836 and the characters it contains. The details of this registry are 1837 specified in [IDNA2008-Tables]. 1839 10.2. IDNA Context Registry 1841 For characters that are defined in the IDNA Character Registry list 1842 as PROTOCOL-VALID but requiring a contextual rule (i.e., the types of 1843 rule described in Section 3.1.2), IANA will create and maintain a 1844 list of approved contextual rules. The details for those rules 1845 appear in [IDNA2008-Tables]. 1847 10.3. IANA Repository of IDN Practices of TLDs 1849 This registry, historically described as the "IANA Language Character 1850 Set Registry" or "IANA Script Registry" (both somewhat misleading 1851 terms) is maintained by IANA at the request of ICANN. It is used to 1852 provide a central documentation repository of the IDN policies used 1853 by top level domain (TLD) registries who volunteer to contribute to 1854 it and is used in conjunction with ICANN Guidelines for IDN use. 1856 It is not an IETF-managed registry and, while the protocol changes 1857 specified here may call for some revisions to the tables, these 1858 specifications have no direct effect on that registry and no IANA 1859 action is required as a result. 1861 11. Security Considerations 1863 11.1. General Security Issues with IDNA 1865 This document in the IDNA2008 series is purely explanatory and 1866 informational and consequently introduces no new security issues. It 1867 would, of course, be a poor idea for someone to try to implement from 1868 it; such an attempt would almost certainly lead to interoperability 1869 problems and might lead to security ones. A discussion of security 1870 issues with IDNA, including some relevant history, appears in 1871 [IDNA2008-Defs]. 1873 12. Acknowledgments 1875 The editor and contributors would like to express their thanks to 1876 those who contributed significant early (pre-WG) review comments, 1877 sometimes accompanied by text, especially Mark Davis, Paul Hoffman, 1878 Simon Josefsson, and Sam Weiler. In addition, some specific ideas 1879 were incorporated from suggestions, text, or comments about sections 1880 that were unclear supplied by Vint Cerf, Frank Ellerman, Michael 1881 Everson, Asmus Freytag, Erik van der Poel, Michel Suignard, and Ken 1882 Whistler, although, as usual, they bear little or no responsibility 1883 for the conclusions the editor and contributors reached after 1884 receiving their suggestions. Thanks are also due to Vint Cerf, 1885 Debbie Garside, and Jefsey Morfin for conversations that led to 1886 considerable improvements in the content of this document. 1888 A meeting was held on 30 January 2008 to attempt to reconcile 1889 differences in perspective and terminology about this set of 1890 specifications between the design team and members of the Unicode 1891 Technical Consortium. The discussions at and subsequent to that 1892 meeting were very helpful in focusing the issues and in refining the 1893 specifications. The active participants at that meeting were (in 1894 alphabetic order as usual) Harald Alvestrand, Vint Cerf, Tina Dam, 1895 Mark Davis, Lisa Dusseault, Patrik Faltstrom (by telephone), Cary 1896 Karp, John Klensin, Warren Kumari, Lisa Moore, Erik van der Poel, 1897 Michel Suignard, and Ken Whistler. We express our thanks to Google 1898 for support of that meeting and to the participants for their 1899 contributions. 1901 Useful comments and text on the WG versions of the draft were 1902 received from many participants in the IETF "IDNABIS" WG and a number 1903 of document changes resulted from mailing list discussions made by 1904 that group. Marcos Sanz provided specific analysis and suggestions 1905 that were exceptionally helpful in refining the text, as did Vint 1906 Cerf, Mark Davis, Martin Duerst, Ken Whistler, and Andrew Sullivan. 1908 13. Contributors 1910 While the listed editor held the pen, this core of this document and 1911 the initial WG version represents the joint work and conclusions of 1912 an ad hoc design team consisting of the editor and, in alphabetic 1913 order, Harald Alvestrand, Tina Dam, Patrik Faltstrom, and Cary Karp. 1914 In addition, there were many specific contributions and helpful 1915 comments from those listed in the Acknowledgments section and others 1916 who have contributed to the development and use of the IDNA 1917 protocols. 1919 14. References 1921 14.1. Normative References 1923 [ASCII] American National Standards Institute (formerly United 1924 States of America Standards Institute), "USA Code for 1925 Information Interchange", ANSI X3.4-1968, 1968. 1927 ANSI X3.4-1968 has been replaced by newer versions with 1928 slight modifications, but the 1968 version remains 1929 definitive for the Internet. 1931 [IDNA2008-Bidi] 1932 Alvestrand, H. and C. Karp, "An updated IDNA criterion for 1933 right to left scripts", July 2008, . 1936 [IDNA2008-Defs] 1937 Klensin, J., "Internationalized Domain Names for 1938 Applications (IDNA): Definitions and Document Framework", 1939 November 2008, . 1942 [IDNA2008-Protocol] 1943 Klensin, J., "Internationalized Domain Names in 1944 Applications (IDNA): Protocol", November 2008, . 1947 [IDNA2008-Tables] 1948 Faltstrom, P., "The Unicode Code Points and IDNA", 1949 July 2008, . 1952 A version of this document is available in HTML format at 1953 http://stupid.domain.name/idnabis/ 1954 draft-ietf-idnabis-tables-02.html 1956 [RFC3490] Faltstrom, P., Hoffman, P., and A. Costello, 1957 "Internationalizing Domain Names in Applications (IDNA)", 1958 RFC 3490, March 2003. 1960 [RFC3492] Costello, A., "Punycode: A Bootstring encoding of Unicode 1961 for Internationalized Domain Names in Applications 1962 (IDNA)", RFC 3492, March 2003. 1964 [Unicode-UAX15] 1965 The Unicode Consortium, "Unicode Standard Annex #15: 1966 Unicode Normalization Forms", March 2008, 1967 . 1969 [Unicode51] 1970 The Unicode Consortium, "The Unicode Standard, Version 1971 5.1.0", 2008. 1973 defined by: The Unicode Standard, Version 5.0, Boston, MA, 1974 Addison-Wesley, 2007, ISBN 0-321-48091-0, as amended by 1975 Unicode 5.1.0 1976 (http://www.unicode.org/versions/Unicode5.1.0/). 1978 14.2. Informative References 1980 [BIG5] Institute for Information Industry of Taiwan, "Computer 1981 Chinese Glyph and Character Code Mapping Table, Technical 1982 Report C-26", 1984. 1984 There are several forms and variations and a closely- 1985 related standard, CNS 11643. See the discussion in 1986 Chapter 3 of Lunde, K., CJKV Information Processing, 1987 O'Reilly & Associates, 1999 1989 [GB18030] "Chinese National Standard GB 18030-2000: Information 1990 Technology -- Chinese ideograms coded character set for 1991 information interchange -- Extension for the basic set.", 1992 2000. 1994 [RFC0810] Feinler, E., Harrenstien, K., Su, Z., and V. White, "DoD 1995 Internet host table specification", RFC 810, March 1982. 1997 [RFC0952] Harrenstien, K., Stahl, M., and E. Feinler, "DoD Internet 1998 host table specification", RFC 952, October 1985. 2000 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 2001 STD 13, RFC 1034, November 1987. 2003 [RFC1035] Mockapetris, P., "Domain names - implementation and 2004 specification", STD 13, RFC 1035, November 1987. 2006 [RFC1123] Braden, R., "Requirements for Internet Hosts - Application 2007 and Support", STD 3, RFC 1123, October 1989. 2009 [RFC2136] Vixie, P., Thomson, S., Rekhter, Y., and J. Bound, 2010 "Dynamic Updates in the Domain Name System (DNS UPDATE)", 2011 RFC 2136, April 1997. 2013 [RFC2181] Elz, R. and R. Bush, "Clarifications to the DNS 2014 Specification", RFC 2181, July 1997. 2016 [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and 2017 Languages", BCP 18, RFC 2277, January 1998. 2019 [RFC2535] Eastlake, D., "Domain Name System Security Extensions", 2020 RFC 2535, March 1999. 2022 [RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", 2023 RFC 2671, August 1999. 2025 [RFC2673] Crawford, M., "Binary Labels in the Domain Name System", 2026 RFC 2673, August 1999. 2028 [RFC2782] Gulbrandsen, A., Vixie, P., and L. Esibov, "A DNS RR for 2029 specifying the location of services (DNS SRV)", RFC 2782, 2030 February 2000. 2032 [RFC3454] Hoffman, P. and M. Blanchet, "Preparation of 2033 Internationalized Strings ("stringprep")", RFC 3454, 2034 December 2002. 2036 [RFC3491] Hoffman, P. and M. Blanchet, "Nameprep: A Stringprep 2037 Profile for Internationalized Domain Names (IDN)", 2038 RFC 3491, March 2003. 2040 [RFC3743] Konishi, K., Huang, K., Qian, H., and Y. Ko, "Joint 2041 Engineering Team (JET) Guidelines for Internationalized 2042 Domain Names (IDN) Registration and Administration for 2043 Chinese, Japanese, and Korean", RFC 3743, April 2004. 2045 [RFC3987] Duerst, M. and M. Suignard, "Internationalized Resource 2046 Identifiers (IRIs)", RFC 3987, January 2005. 2048 [RFC4290] Klensin, J., "Suggested Practices for Registration of 2049 Internationalized Domain Names (IDN)", RFC 4290, 2050 December 2005. 2052 [RFC4690] Klensin, J., Faltstrom, P., Karp, C., and IAB, "Review and 2053 Recommendations for Internationalized Domain Names 2054 (IDNs)", RFC 4690, September 2006. 2056 [RFC4713] Lee, X., Mao, W., Chen, E., Hsu, N., and J. Klensin, 2057 "Registration and Administration Recommendations for 2058 Chinese Domain Names", RFC 4713, October 2006. 2060 [Unicode-Security] 2061 The Unicode Consortium, "Unicode Technical Standard #39: 2062 Unicode Security Mechanisms", August 2008, 2063 . 2065 [Unicode-UAX31] 2066 The Unicode Consortium, "Unicode Standard Annex #31: 2067 Unicode Identifier and Pattern Syntax", March 2008, 2068 . 2070 [Unicode-UTR36] 2071 The Unicode Consortium, "Unicode Technical Report #36: 2072 Unicode Security Considerations", July 2008, 2073 . 2075 Appendix A. Change Log 2077 [[ RFC Editor: Please remove this appendix. ]] 2079 A.1. Changes between Version -00 and Version -01 of 2080 draft-ietf-idnabis-rationale 2082 o Clarified the U-label definition to note that U-labels must 2083 contain at least one non-ASCII character. Also clarified the 2084 relationship among label types. 2086 o Rewrote the discussion of Labels in Registration (Section 7.1.2) 2087 and related text about IDNA-validity (in the "Defs" document as of 2088 -04 of this one) to narrow its focus and remove more general 2089 restrictions. Added a temporary note in line to explain the 2090 situation. 2092 o Changed the "IDNA uses Unicode" statement to focus on 2093 compatibility with IDNA2003 and avoid more general or 2094 controversial assertions. 2096 o Added a discussion of examples to Section 7.1 2098 o Made a number of other small editorial changes and corrections 2099 suggested by Mark Davis. 2101 o Added several more discussion anchors and notes and expanded or 2102 updated some existing ones. 2104 A.2. Version -02 2106 o Trimmed change log, removing information about pre-WG drafts. 2108 o Adjusted discussion of Contextual Rules to match the new location 2109 of the tables and some conceptual material. 2111 o Rewrote the material on preprocessing somewhat. 2113 o Moved the material about relationships with IDNA2003 to be part of 2114 a single section on transitions. 2116 o Removed several placeholders and made editorial changes in 2117 accordance with decisions made at IETF 72 in Dublin and not 2118 disputed on the mailing list. 2120 A.3. Version -03 2122 This special update to the Rationale document is intended to try to 2123 get the discussion of what is normative or not under control. While 2124 the IETF does not normally annotate individual sections of documents 2125 with whether they are normative or not, concerns that we don't know 2126 which is which, claims that some material is normative that would be 2127 problematic if so classified, etc., argue that we should at least be 2128 able to have a clear discussion on the subject. 2130 Two annotations have been applied to sections that might reasonably 2131 be considered normative. One annotation is based on the list of 2132 sections in Mark Davis's note of 29 September (http:// 2133 www.alvestrand.no/pipermail/idna-update/2008-September/002667.html). 2135 The other is based on an elaboration of John Klensin's response on 7 2136 October (http://www.alvestrand.no/pipermail/idna-update/2008-October/ 2137 002691.html). These should just be considered two suggestions to 2138 illuminate and, one hopes, advance the Working Group's discussions. 2140 Some additional editorial changes have been made, but they are 2141 basically trivial. In the editor's judgment, it is not possible to 2142 make significantly more progress with this document until the matter 2143 of document organization is settled. 2145 A.4. Version -04 2147 o Definitional and other normative material moved to new document 2148 (draft-ietf-idnabis-defs). Version -03 annotations removed. 2150 o Material on differences between IDNA2003 and IDNA2008 moved to an 2151 appendix in Protocol. 2153 o Material left over from the origins of this document as a 2154 preliminary proposal has been removed or rewritten. 2156 o Changes made to reflect consensus call results, including removing 2157 several placeholder notes for discussion. 2159 o Added more material, including discussion of historic scripts, to 2160 Section 3.2 on registration policies. 2162 o Added a new section (Section 7.2) to contain specific discussion 2163 of handling of characters that are interpreted differently in 2164 input to IDNA2003 and 2008. 2166 o Some material, including this section/appendix, rearranged. 2168 A.5. Version -05 2170 o Many small editorial changes, including changes to eliminate the 2171 last vestiges of what appeared to be 2119 language (upper-case 2172 MUST, SHOULD, or MAY) and small adjustments to terminology. 2174 A.6. Version -06 2176 o Removed Security Considerations material and pointed to Defs, 2177 where it now appears as of version 05. 2179 o Started changing uses of "IDNA2008" in running text to "in these 2180 specifications" or the equivalent. These documents are titled 2181 simply "IDNA"; once they are standardized, "the current version" 2182 may be a more appropriate reference than one containing a year. 2184 As discussed on the mailing list, we can and should discuss how to 2185 refer to these documents at an appropriate time (e.g., when we 2186 know when we will be finished) but, in the interim, it seems 2187 appropriate to simply start getting rid of the version-specific 2188 terminology where it can naturally be removed. 2190 o Additional discussion of mappings, etc., especially for case- 2191 sensitivity. 2193 o Clarified relationship to base DNS specifications. 2195 o Consolidated discussion of lookup of unassigned characters. 2197 o More editorial fine-tuning. 2199 A.7. Version -07 2201 o Revised terminology by adding terms: NR-LDH-label, Invalid-A-label 2202 (or False-A-label), R-LDH-label, valid IDNA-label in 2203 Section 1.3.3. 2205 o Moved the "name server considerations" material to this document 2206 from Protocol because it is non-normative and not part of the 2207 protocol itself. 2209 o To improve clarity, redid discussion of the reasons why looking up 2210 unassigned code points is prohibited. 2212 o Editorial and other non-substantive corrections to reflect earlier 2213 errors as well as new definitions and terminology. 2215 A.8. Version -08 2217 o Slight revision to "contextual" discussion (Section 3.1.2) and 2218 moving it to a separate subsection, rather than under "PVALID", 2219 for better parallelism with Tables. Also reflected Mark's 2220 comments about the limitations of the approach. 2222 o Added placeholder notes as reminders of where references to the 2223 other documents need Section numbers. More of these will be added 2224 as needed (feel free to identify relevant places), but the actual 2225 section numbers will not be inserted until the documents are 2226 completely stable, i.e., on their way to the RFC Editor. 2228 Author's Address 2230 John C Klensin 2231 1770 Massachusetts Ave, Ste 322 2232 Cambridge, MA 02140 2233 USA 2235 Phone: +1 617 245 1457 2236 Email: john+ietf@jck.com