The real issue: interopability, and a proposal (Was: Consensus Call on Latin Sharp S and Greek Final Sigma)

John C Klensin klensin at
Tue Dec 1 12:24:28 CET 2009

--On Tuesday, December 01, 2009 10:14 +0000 Shawn Steele
<Shawn.Steele at> wrote:

>> There seems to be little existing deployment of ß-labels out
>> there, at least on the web 
> I would guess that be a case against making eszett PVALID.
> Existing users needing Eszett are instead making due with ss,
> so adding Eszett explicitly would put those users at risk.
> It'd be better to enable them to have a display form so they
> could indicate that Eszett was their preferred form.


Please remember that, for many years, existing users needing ö
(o-umlaut in the German context) used "oe", so carrying this
argument very far would rapidly turn into an argument against
any IDNs involving decorated Latin characters.  (The arguments
for IDNs based on non-Latin scripts would remain intact.).  The
situation with existing registrations involving "oe" would be
almost the same as the Eszett/ "ss" situation -- one could not
tell mechanically whether the non-ASCII form was actually

Long after registrations involving "ö" started to be permitted,
many users continued to use the "oe" forms.  The mapping from
one to the other was never, as far as I know, embedded in code,
but, from a user standpoint, the situation is much the same:
some strings containing "oe" are interchangeable with the
corresponding strings with "ö" and, if both are not registered
to the same entity and either bundled or redirected, the user
somehow needs to keep track of what is registered and pick the
correct name.

And someone guessing the domain name for Herr Möller would have
had to know to guess "moeller" before IDNs were introduced and
which one to try now that they are available. 

"ß" isn't a display form for "ss" any more than "ö" is a
display form for "oe".  They are different characters.  One
could plausibly consider them display forms if "ß" could be
substituted for every instance of "ss" or "ö" could reasonable
be substituted for every instance of "oe", but that is not the
case: Göthestraße is just wrong, not a display variation of
Goethestrasse.  As Patrik points out, the substitutions (both of
them) are _always_ wrong in Swedish.

I note, fwiw, that many registries and users seem to have
survived the introduction of "ö".  I conclude that the
transition problem isn't quite as problematic as you and some
others are claiming.    That doesn't mean it isn't a problem,
but perhaps we are getting a little carried away by hyperbole.



More information about the Idna-update mailing list