I've done so minor edits to this based on the discussion on the mailing list plus some other comments I've received.
Erik 1.1 Problem Statement The IDNA specification solves the problem of extending the repertoire of characters that can be used in domain names to include the Unicode repertoire (with some restrictions). IDNA does not extend the service offered by DNS to the applications. Instead, the applications (and, by implication, the users) continue to see an exact-match lookup service. Either there is a single exactly-matching name or there is no match. This model has served the existing applications well, but it requires, with or without internationalized domain names, that users know the exact spelling of the domain names that the users type into applications such as web browsers and mail user agents. The introduction of the larger repertoire of characters potentially makes the set of misspellings larger, especially given that in some cases the same appearance, for example on a business card, might visually match several Unicode code points or several sequences of code points. IDNA allows the graceful introduction of IDNs not only by avoiding upgrades to existing infrastructure (such as DNS servers and mail transport agents), but also by allowing some rudimentary use of IDNs in applications by using the ASCII representation of the non-ASCII name labels. While such names are very user-unfriendly to read and type, and hence are not suitable for user input, they allow (for instance) replying to email and clicking on URLs even though the domain name displayed is incomprehensible to the user. In order to allow user-friendly input and output of the IDNs, the applications need to be modified to conform to this specification. IDNA uses the Unicode character repertoire, which avoids the significant delays that would be inherent in waiting for a different and specific character set be defined for IDN purposes by some other standards developing organization. 1.2 Limitations of IDNA <EXISTING section 6.6 moved to here> The IDNA protocol does not solve all linguistic issues with users inputting names in different scripts. Many important language-based and script-based mappings are not covered in IDNA and need to be handled outside the protocol. For example, names that are entered in a mix of traditional and simplified Chinese characters will not be mapped to a single canonical name. Another example is Scandinavian names that are entered with U+00F6 (LATIN SMALL LETTER O WITH DIAERESIS) will not be mapped to U+00F8 (LATIN SMALL LETTER O WITH STROKE). <ADDED> An example of an important issue that is not considered in detail in IDNA is how to provide a high probability that a user who is entering a domain name based on visual information (such as from a business card or billboard) or aural information (such as from a telephone or radio) would correctly enter the IDN. Similar issues issues exist for ASCII domain names, for example the possible visual confusion between the letter 'O' and the digit zero, but the introduction of the larger repertoire of characters creates more opportunities of similar looking and similar sounding names. Note that this is a complex issue relating to languages, input methods on computers, and so on. Furthermore, the kind of matching and searching necessary for a high probability of success would not fit the role of the DNS and its exact matching function. ---
