Mixing scripts (Unicode versions (Criteria for exceptional characters))

[Erik van der Poel]

Sarcasm noted, but if I may comment on the users being "desensitized
to the display of raw Punycode", instead of my idea of showing a
shaded rectangle, about the size of the original label, an application
could choose to display the little boxes that IE6 has been displaying
for IDNs for a while (which means that some users may have seen
those). If the user _still_ thinks that that site is his bank, then no
amount of education will help him. Such a tactic would also thwart
those who attempt to use the appearance of the raw Punycode to trick
the user (e.g. xn--wellsfargo).

This tactic could be used by any application that currently chooses to
display Punycode when the Unicode version is somehow deemed dangerous,
whether the implementors are in the disallow-mixed-scripts camp or
not. Cut-n-paste shouldn't be a problem either; applications have been
able to cut or copy text that does not resemble the selected object
for quite some time now.

Erik

On 12/21/06, Cary Karp <ck at nic.museum> wrote:
> If we can't make that happen, then I suppose anything will be fair game,
> including the lucrative market that can be established to serve a user
> community which has been desensitized to the display of raw Punycode.
> We've already got a pretty good sense of the scope of the market
> potential for xn--payload labels, and I suspect we're not far from
> discovering Unicode/Punycode pairs that could credibly be brandable in
> both forms.