Re: [idn] Requirements: De-specifying Unicode

[ned.freed@innosoft.com]

> Paragraph 3 of section 2.2 says:

> >In order to simplify the requirements, it is assumed that characters
> >from the Unicode character set will be used for the protocol on the
> >wire. The protocol may use multiple character sets if doing so meets
> >all the other requirements in this document. However, this should
> >not constrain which character set or sets an IDN implementation may
> >use for its user interface, or for the storage of records in a master
> >file.

> While I personally agree with the first sentence, I don't think it is an
> actual requirement. Some people have proposed multiple character sets with
> character set tagging. I suggest we remove the first sentence in order to
> allow more proposals.

I sure hate to do this, but I reluctantly have to agree with Paul that it is
probably best to leave the door open to tagging proposals. Although I don't
see how such things can possibilty meet the canonicalization requirements
any viable IDN scheme is bound to have...

				Ned