Comments on the ccTLD IDN FT program

[Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>]

Tina,

As I can't tell if there will be an extension of public comments, this 
hurried note will have to do.
The ccTLD IDN FT was broken in two important ways ab initio. First, 
the one script per presumption, which became a rule without sufficient 
reflection, and second, the presumption that only non-Latin scripts 
are necessary, which also became a rule without sufficient reflection.
The one-script-per rule is sensible for monolingual countries, and 
either comedic or potentially tragic for countries with either a 
healthy linguistic plurality, or in which linguistic hegemony is 
precarious and maintained by force. There are several candidates for 
either, and I'll just point to the canonical example of 11 official 
scripts in India, for 22 official languages.
The non-Latin rule is also sensible for countries in which the 
expansion of Western European literary culture in the 16th to 20th 
centuries was not pervasive -- the Middle East and North Africa, 
Eastern Europe, and Asia. Where cultures adapted to Latin script 
hegemony, the Americas, Oceania, Sub-Saharan Africa, the non-Latin 
rule eliminates the adaptations, usually using diacriticals, which are 
encoded as Extended Latin in Unicode, of non-Latin languages which use 
Latin Script.
So this point is not lost, on restricted keyboards, even Arabic has a 
Latin keyboarding convention used by millions, on ordinary keyboards, 
Cherokee, Cree, and other unrelated indigenous languages with 
non-Latin scripts are as commonly keyboarded in Latin as Arabic.
What we call "decorated Latin" has been excluded by the ccTLD IDN FT, 
which has the effect of limiting the benefit of the ccTLD IDN FT to 
the majority populations of East Asian, Eastern European, and West 
Asian and North African states, and utterly denying it to the 
pluralities of populations of the Americas who do not ordinarily use a 
Latin language, but who do ordinarily use Latin script, with extensions.
With those two areas covered, two subjects that affect IDN users 
differently -- "variants" and states as the sole authors of 
applications for scripts.
The rhetoric of the ccTLD IDN FT has been crippled by the use of the 
word "variant" for vastly different technical problems. The 
historically tragic "Han Unification" committed by the Unicode 
Consortium has resulted in the unfortunate treatment of Simplified and 
Traditional Chinese characters as "variants", rather than as two 
distinct scripts of tens of thousands of characters sharing several 
thousand characters in common. The less tragic "Arabic Script" 
unification of Arabic and Farsi/Dari has also created a kind of 
variant problem, one that showed up during the IDNAbis work that 
touched on Arabic and Eastern Arabic digits and their directionality 
and similarity. Finally, there are the true variants, arising from the 
context dependent nature of some scripts, such as Arabic, with an 
initial, medial, and terminal form for each of the characters.
Calling these all "variants" has lead to a year long inability to 
express usefully what is meant by equivalences of variants, to the 
consternation of the IETF DNSEXT WG that has attempted to assist ICANN 
in spite of a history of shared mistakes and misunderstandings,
and put some non-trivial additional tensions on the relationship 
between the largest single language group on the planet and ICANN.
The restriction of agency for scripts, and therefore languages, to 
state, has adverse consequences. It is reasonable to offer Yiddish, 
language written both in Latin script and in Hebrew script, by any 
state using a right-to-left bidirectional script. Unfortunately, for 
language policy that does not meet rational scrutiny, no such state is 
interested in the survival of the Yiddish language in Hebrew script. 
Additional examples of state disinterest in important languages exist. 
State elites, whether purely for policy rationals, or for market 
rationals, are unlikely sources of agency to adopt a minority 
language, even if no other "Fast Track" choices are available to the 
state elites.
Yes, ICANN has made some gesture of service towards the Han, Arabic, 
and Cyrillic script users, and will get to some of the Indic script 
users before the end of the FT process. But we could have done much 
more, and the reveiw cannot legitimately be "flawless execution, with 
champers and backslapping all around."
Eric Brunner-Williams
in a personal capacity