UCAS and related scripts
ebw at abenaki.wabanaki.net
Mon Apr 7 19:08:46 CEST 2008
John C Klensin wrote:
> --On Monday, 07 April, 2008 08:56 -0700 Eric Brunner-Williams
> <ebw at abenaki.wabanaki.net> wrote:
>> Thank you for writing back.
>> I'm more interested in the full stop (166e in your table) than
>> in the symbol for a foreign cult (166d in the same text).
>> I'm guessing that the rational for the full stop exclusion is
>> not that its "confusingly similar to" some other character,
>> but because it is functionally equivalent to "dot" (and is
>> always translated into roman script as ".").
>> Which rational, or something I've not guessed, is present?
> There may be some exceptions, for which it is better that you
> check the I-Ds rather than my trying to summarize and get it
> wrong but, in general, we have not disallowed characters because
> of confusing similarity (aka "phishing risk") alone. In the
> case of U+166E, it is excluded because it has the Unicode
> General Category "Po" (other punctuation) and all such
> characters are Disallowed.
> Correctly or incorrectly, U+166D is also in "Po". If I
> correctly understand your comments and those of Michael Everson,
> it might better have been classified as a symbol of some
> flavor, but that would make no difference for our purposes,
> since symbols are also disallowed.
> If you haven't had time to look at the Rationale document
> (draft-klensin-idnabis-issues) yet, the general theme of this
> part of the new approach is to return to the "LDH" rule
> associated with traditional hostnames and then try to generalize
> the "letter" and "digit" parts to corresponding characters in
> Unicode, rather than, e.g., trying to see how much can be
> included without causing problems. If that formulation is
> working out well for indigenous North American writing systems,
> it suggests that the rules we are using generally worked rather
> than because we conducted a character-by-character examination
> and then invented new classifications.
I've gotten through Patrick's draft, and Harald's draft, and a few pages
into your protocol draft, so I haven't made it to your issues draft.
Exclusion of a punctuation character as a punctuation character is
reasonable. I've looked at a bunch of texts, from eastern Algonquin
through (western) Carrier, not pretending that my Cree and Siksika is
either necessary nor sufficient, and I can't see an interesting use case
where an infix full-stop would work as a domain name.
Just a reminder, I speak for myself, not the imaginary union of all
literate and semi-literate present and future consumers and producers of
text written in northern syllabic scripts, who may or may not find the
DNS a source of joy and wonder.
More information about the Idna-update