Re: [I18nrp] Mappings for IDNA2008 ?

[Nico Williams <[e-mail]>]

On Mon, Feb 04, 2019 at 07:25:54PM -0500, John Levine wrote:
> In article <[e-mail]> you write:
> >> Are there any published IDNA2008 mappings?  As far as I can tell,
> >> everyone uses one from UTS46 by default, and it's not very good.
> >>
> >Examples of its badness, please.
> 
> As I understand it:
> 
> If you speak Turkish, the case folding is wrong.

So, IDNA says not to case-fold.  So users need to get case right -- oof.
Now, mostly users are accustomed to seeing domainnames in lower-case,
and type them in that way.  Sadly, mobile UIs stupidly up-case the first
letter (yay!), so if we don't at least down-case that first letter, we
have a problem for IDNs (ASCII labels are matched case-insensitively, so
they're OK; it's only IDNs that have this problem).

Now, if you never up-case, and only down-case, and if the user enters
the i or dotless i in lower-case, then there's no case folding issue for
Turkish.

The Turkish issue is academic, not a real issue.

Now, IF we could get all the broken mobile URI entry UIs in the world to
stop up-casing the first letters domainnames, then we could just train
users to enter domainnames in lower-case (or whatever case mix is
appropriate) and we'd have no problem...

...except users are not going to remember correct mix-case domainnames,
so case folding is basically necessary.  That there are a few ambiguous
cases like the Turkish dotless i seems like one of those "tough, deal
with it" moments.

Nico
--