Mixing scripts

[Soobok Lee]

On Sat, Dec 23, 2006 at 12:03:52AM +0100, Kent Karlsson wrote:
> 
> Paul Hoffman wrote:
> > Mostly true, but there is an important exception. Nameprep section 6 
> > (bidi checking) works on the whole label, not just on individual 
> > characters. Fortunately, the rule defined there (inherited from 
> > Stringprep) is extremely simple to describe and therefore to 
> > implement.
> 
> The "all the chars from the same script (plus common and inherited)"
> is just as simple a rule, more correct, and prevents more spoofs.
> So indeed the same-scripts rule, as explained by Mark, is a
> generalisation and correction of this already existing (but somewhat
> flawed) rule. It needs an additional slight modification allowing
> the "Japn" mix (in one "label"), but IFAIK there is no particular
> need for any other similar mix (not even for Korean, for instance).

If you visit http://www.lgtelecom.com/ (#3 mobile telco in korea), 
please look into the title bar which displays "LG(telecom in hangul)".
In Korea, latin+hangul is common, while CJK+hangul is rare.
I can't exclude even the need for CJK+hangul label in Korea from those
whose like archaic and peculiar homepage address.
For CJK+hangul, confusibility risk is low, even though its typing is more 
difficult than that of pure hangul label.

Soobok

> 
> So it is certaily not a big leap to put this in "the protocol" (in
> "stringprep" actually). A similar (though more limited and flawed)
> rule already exists.
> 
> I would add that the first character should be a base character,
> and some "common" characters should be omitted, but that are
> other unrelated rules.
> 
> 	/kent k
> 
> _______________________________________________
> Idna-update mailing list
> Idna-update at alvestrand.no
> http://www.alvestrand.no/mailman/listinfo/idna-update