Re: [idn] case preservation<20011009004803.28618.qmail@cr.yp.to><00b501c15078$afb77ce0$ec1bd9d2@temp><20011009043734.15576.qmail@cr.yp.to><58648759.1002614060@localhost>

[Martin Duerst <duerst@w3.org>]

At 06:51 01/10/09 +0000, D. J. Bernstein wrote:
>If you're not going to do a thorough job of nameprep, then you shouldn't
>be trying to deceive users into believing that IDNs can be safely typed
>from textual printouts.

I think it was never a goal to have mixed-script names
(e.g. <A-Greek><O-Cyrillic><L-Latin>) to be easy to type.
The main problem in such cases for actual input is that
you have to switch (software) keyboards, even if it is
very clear which script the character comes from.
There was never even a goal to fully support completely
mixed-script labels, because they don't really make sense.
However, because names e.g. of the form latin-greek
could make sense, mixed-script labels as such were
included in the requirements.


>It would be better to omit nameprep entirely;
>then users will stick to mechanisms that work, such as clicking on URLs.
>(Bad characters should still be avoided in registrations, of course.)

Bad character combinations should be avoided with registration.
The only registrations for things like <A-Greek><O-Cyrillic><L-Latin>
will come from spoofers, and the only way such spoofers have
a chance is through clicking on URIs or a similar interface.


Regards,    Martin.