[idn] IDN Implementation Status Update

Gervase Markham <gerv@mozilla.org> Fri, 22 July 2005 16:20 UTC

Message-ID: <42E11B09.6030203@mozilla.org>
Date: Fri, 22 Jul 2005 17:12:57 +0100
From: Gervase Markham <gerv@mozilla.org>
Organization: mozilla.org
User-Agent: Mozilla Thunderbird 1.0 (X11/20041206)
MIME-Version: 1.0
To: iab-idn@lists.paf.se, mb-secissues@opera.com, idn@ops.ietf.org
Subject: [idn] IDN Implementation Status Update
Content-Type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-Transfer-Encoding: 7bit
Sender: owner-idn@ops.ietf.org
Precedence: bulk
Content-Transfer-Encoding: 7bit

This is a message to tell interested parties the current status of the 
mozilla.org changes to Firefox regarding IDN.

We have implemented a TLD whitelist system, which currently contains 21 
TLDs for which we correctly display IDN domain names in the UI.
http://www.mozilla.org/projects/security/tld-idn-policy-list.html
Any IDN domain name in a non-whitelisted TLD displays as punycode. This 
is a security feature and so there is no user interface for adding or 
removing TLDs.

Any registry which wishes to be added to the whitelist should follow the 
instructions on that page. In terms of what constitutes a homograph, we 
are being guided by the Unicode Consortium's confusables list:
http://www.unicode.org/draft/reports/tr36/data/confusables.txt
and by common sense. Our policy in this area is still somewhat in flux - 
in particular, we are not yet sure whether we should require that 
registries to consider two characters which differ only in accent 
(sometimes by the shade of a single pixel at normal font sizes) as 
homographic. In the mean time, we strongly advise that registries do this.

We have implemented a character blacklist, which will soon contain 
'DIVISION SLASH' (U+2215) and 'FRACTION SLASH' (U+2044). After that, we 
may extend it to forbid more characters which may be used to spoof URL 
punctuation.
https://bugzilla.mozilla.org/show_bug.cgi?id=301694
This is not meant to prejudice the outcome of the current IAB-IDN 
discussions on potentially reducing the number of characters permitted 
in IDN, but we feel the danger posed by the use of such characters in 
3rd and 4th level domains is great enough to require an immediate ban. 
Any domain name which contains one or more of these characters displays 
as punycode.

We wish to thank Opera Software for their help in creating the initial 
whitelist and providing suggestions for the character blacklist.

Please note that I will be away from Saturday 23rd of June until Monday 
8th August, so please do not expect any email replies during that time. :-)

Gerv

[idn] IDN Implementation Status Update Gervase Markham