draft-klensin-idnabis-protocol-04 section 4.5
Erik van der Poel
erikv at google.com
Thu Mar 27 15:50:29 CET 2008
Simon, the IDNA200X drafts do not include the "mapping" (case-folding,
NFKC and "map to nothing" (delete)) that was included in IDNA2003.
The consensus so far is that all of that "mapping" is essentially
moved out of the core 4 specs. One mapping spec is the one that Mark
produced, and we are now calling it "pre-processing". I believe you
already received a pointer to that draft.
Another level of mapping is described in the early steps of
registration and lookup in the IDNA200X protocol draft. This is the UI
level, where apps are, for example, free to map the dotted and dotless
letters 'i' the Turkish way if the user has asked the app to assume
Turkish conventions. This area is somewhat controversial, since a
proliferation of language-specific mappings may create some confusion.
Nevertheless, some app developers may consider this essential for a
good user experience.
So the bottom line is that the current four IDNA200X drafts only
specify what is allowed at the lowest level(s). The higher levels,
such as HTML, UI and so on, are to be specified in separate specs.
On Thu, Mar 27, 2008 at 6:02 AM, Harald Alvestrand <harald at alvestrand.no> wrote:
> Simon Josefsson wrote:
> > Harald Alvestrand <harald at alvestrand.no> writes:
> >>> What I'm trying to understand is what an IDNA200x implementation will do
> >>> (i.e., which output string or what error) when the user types 'josefßon'
> >>> or 'dªtªkonsult'.
> >> Read the drafts. It helps.
> > My conclusion from reading the drafts is that typing either string will
> > lead to an error state in the application.
> > In IDNA2003 the strings are successfully looked up in DNS.
> > If this interpretation is incorrect, please clarify.
> I agree with your interpretation. So I guess the draft is clear enough.
> Idna-update mailing list
> Idna-update at alvestrand.no
More information about the Idna-update