[Latingp] Latin Small Letter Sharp S (ß) U+00DF

Mats Dufberg mats.dufberg at internetstiftelsen.se
Thu Aug 22 13:46:08 UTC 2019


It seems like we have to accept that the IDNA 2003 behavior will stay for quite some time. Besides Chrome and IE, which predates the release of IDNA 2008, even Microsoft Edge, which is a new web browser has the IDNA 2003 behavior on Sharp S.

Users of those browsers will not be able to reach domain names under a TLD with sharp S unless there is an activated variant with double S. I really think that is a significant problem. I think we agree that the behavior of those browsers is unwanted, but I am not sure that we should consider it to be a bug that will be fixed.


Mats

---
Mats Dufberg
DNS Specialist
Internetstiftelsen (The Swedish Internet Foundation)
Mobile: +46 73 065 3899
https://internetstiftelsen.se/


From: Latingp <latingp-bounces at icann.org> on behalf of Bill Jouris <bill.jouris at insidethestack.com>
Reply to: Bill Jouris <bill.jouris at insidethestack.com>
Date: Monday, 19 August 2019 at 15:32
To: ICANN Latin GP <latingp at icann.org>
Subject: [Latingp] Latin Small Letter Sharp S (ß) U+00DF

Dear colleagues,

I had a couple of thoughts over the weekend.

First, I think that in order to make progress we need to unpack our analysis a little.  In particular, we need to consider each option, for both misconnection and failure of connection, separately for IDNA 2003 and for IDNA 2008.  Then we will have the information in front of us to decide how much of a problem we have.

For example, consider Option 2 (Sharp S and double S as blocked variants).  (For this discussion, assume a TLD using the Sharp S exists.  Thus the TLD which only differs by having a double S is blocked.)  If the user enters a valid domain name using the Sharp S, on a browser which uses IDNA 2003, the browser converts it to double s before going to the DNS.  The DNS doesn't find the name, and so the connection fails.  Whereas a browser with IDNA 2008 goes to the DNS with the correct name, and the connection succeeds.

Does this constitute a change of behavior?  Absolutely.  But is this change, "instability" if you will, a significant problem?  I would suggest that it is not.  To call it a significant problem is to argue that bugs should not be fixed.  A bug fix, after all, necessarily results in a change of behavior; that is its whole purpose. Back-level software (which IDNA 2003 is) will always give problematic results.  That's why we upgrade.

Certainly an ideal solution will have no misconnection under either level of IDNA, and no failed connection when the user enters a valid domain name.  But unless we find such a paragon, we will need to find the "least bad" solution.  And having this additional data in front of us will help with that.


Second, I want to comment on the Option for Sharp S and double S with no variant relationship.  I have a question for those who are more fluent in German than I -- which is most of you.  Are there any examples of two different German words which are identical except that one is rendered with a Sharp S and the other is rendered strictly with a double S?

It seems to me that, unless there are, a "not variant" judgement is unsustainable and the option should be rejected out of hand.  Regardless of what this analysis might find.


Bill Jouris
Inside Products
bill.jouris at insidethestack.com
831-659-8360
925-855-9512 (direct)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/latingp/attachments/20190822/8e77a0c9/attachment.html>


More information about the Latingp mailing list