----- Original Message ----- From: "Paul Hoffman / IMC" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Wednesday, October 31, 2001 1:46 PM Subject: Re: [idn] Re: stringprep and unassigned code points
> At 12:58 PM +0900 10/31/01, Soobok Lee wrote: > >People are familiar with case-insensitiveness in hostnames in URL. > >They often type in WWW.AOL.COM with CAPSLOCK ON inadvertantly even > >after they saw lowercased advertized hostnames and > >have been successful so far in reaching the right webhost for > >LDH ones and will be so even for future Cyrillic IDN ones. > > How is this statement at all relevant to the topic being discussed? > Are there any scripts that are being encoded in future versions of > ISO 10646 that will be affected by CAPS LOCK? Switching cases in Greek/Cyrillic IME need CAPS LOCK. Don't they? New bicameral script IME's behavior on CAPS LOCK may resemble them as a natural choice. > > Please note that the discussion we are having is about future > characters and scripts being encoded *that also have nameprep > mappings to other characters*. If future characters are included > without any mappings, everything works just as the user would expect. X and Y always look identical in the case of NFC X->Y . And, script IME (or platform) often has its own preferred form between X or Y. That's often beyond the choice and capability of the end user. I am not sure how many NFC will be defined for new script. But that theory will work if any. Soobok Lee > > What Yves brought up is only valid if there are mappings; that will > have to be explained better in the stringprep document. > > --Paul Hoffman, Director > --Internet Mail Consortium >
