Re: [idn] time to move

[Keith Moore <moore@cs.utk.edu>]

> At 00:16 01/05/24 -0400, Keith Moore wrote:
> > > I don't have a whole lot of 3Mb/sec Type I Ethernet lying about these
> > > days (see above).
> >
> >perhaps not.  but the folks I know where were using IP over 3Mb/sec
> >Ethernet are still using IP.  IPv4 is starting to look obsolete due
> >to a shortage of addresses, but I don't know anyone who thinks it's
> >going to be discarded anytime soon.
> 
> Yes. But if that example applies to anything, it applies with
> IP4 <=> ACE, IP6 <=> UTF-8, i.e. once we have ACE, we will have
> big problems to get to UTF-8.

Actually, we will have big problems to "get to" UTF-8 whether we
go to ACE first or not.    I'd argue that it's easier to "get to"
UTF-8 by going to ACE first because going to ACE allows folks to
get benefit without having to upgrade every single component
in the signal path. 

Of course, the goal of IDN isn't to "get to" utf-8 at all; it is 
to allow domain names to use characters outside of the ASCII repertoire.
If you want to "get to" utf-8 you're trying to solve a different
problem. If you insist that that problem be piggybacked on the
IDN problem then you're trying to make the IDN problem more difficult
than it needs to be.
 
> > > Isn't the line of march of uniqueness and consistency and so forth 
> > predicated
> > > upon the inexistence of ... "island(s)", and the non-deployment of utf8?
> >
> >it's predicated on the non-deployment of utf-8 *in Internet applications*
> >*at the time the IDN solution is deployed*.
> 
> utf-8 is well supported by the recent generations of browsers.

Browsers are but a small subset of internet applications.

> The situation is similar for e-mail, with some exceptions.

Email support for utf-8 is essentially nonexistent.

> Speaking about 'non-deployment of utf-8 in Internet applications'
> is just spreading bad rumors. 

You might not like it, but it's the truth.

> If a bit more people had had the
> guts to go for utf-8 a few months ago, we would already have quite
> some things running, end-to-end.

we would have some prototype code for each of the components.
that doesn't mean it would be deployable in the real world.

no application upgrade in the history of the internet has had 
signifcant deployment in a few months' time.  not even close.
even the MIME upgrade, which was far less disruptive than
what you are proposing, took many years.

we have prototype code for ACE solutions also.  ability to develop
a prototype isn't the issue here.

> >it is currently the case
> >that few Internet applications support utf-8, and this isn't likely to
> >change in the near future.
> 
> Can you please help me understand your facts by telling me how you
> count?

I examine every internet application to which I have easy access - which is
to say, several dozen.  I see how many of them would work with UTF-8 domain
names.  The answer is zero.  Even the applications that accept UTF-8 input
don't do nameprep, neither do they  know better than to pass UTF-8 names 
to other applications that don't deal with UTF-8.  

the "just use UTF-8 in existing applications" approach has always been
a non-starter.
 
> >if you want to wait until everything else  is using utf-8 before deploying
> >IDNs that use utf-8, feel free.  I suspect that most of us don't want
> >to wait that long.  I also suspect that the fastest way to get to having
> >applications use utf-8 natively (if that's even worth the trouble)
> >is to use ACE IDNs in the near term.
> 
> I think this is very seriously wrong. For a very large number of applications,
> the only thing you have to do to work with utf-8 is make sure they are
> 8-bit compatible. In many cases, that's already the case. Even for something
> as old as sendmail, an easy hack to achieve this has been described on this 
> list.

this analysis is simply incorrect, as numerous people have already pointed out.

this discussion can serve no further purpose.  I can't make you believe 
something that you don't want to believe, and I won't waste my energy trying.
but there is no basis for agreement between us, because your assessment of 
the situation is completely incompatible with mine.

Keith