ICANN ICANN Email List Archives


<<< Chronological Index >>>    <<< Thread Index >>>

[no subject]

  • To: <idn-variant-tld@xxxxxxxxx>
  • Subject:
  • From: "Dr. Sarmad Hussain" <sarmad@xxxxxxxxxx>
  • Date: Tue, 5 Apr 2011 00:02:23 +0500

A few comments in my personal capacity:


1.        It should be clarified in the document if the issues to be
identified are relevant at language level or script level (or both levels).
For example, two characters may be variants at language level, but may not
be variants at script level.  Language level requires a much larger exercise
as there may be many languages written using a single script.  Language
level issues are normally relevant for ccTLDs and script level issues could
be relevant for gTLDs.  The current document refers to both types of TLDs
but does not make the distinction of language vs. script.  


2.       The five choices are good as they cover various types of scripts
(logographic, abjad, abugida, and alphabet).  However, though Chinese,
Arabic, Cyrillic and Latin represent scripts (e.g. explicitly listed as such
in Unicode (see http://www.unicode.org/charts)), "Indic" does not represent
a single script but may refer to a larger set of scripts depending on how it
is interpreted.  Linguistically "Indic" may be used to refer to Brahmic
derived scripts; then it includes a wide variety of scripts (including
Tibetan, Devanagari, Gurmukhi, Bangla, Tamil, Sinhala, Lao, Thai, Khmer,
Burmese, Javanese, Tagalog, and many more), which are linguistically too
diverse for a small sub-group of to analyze in a short amount of time.
Geographically "Indic" may be used to refer to Indian scripts (e.g. see
Figure 1 on Page 8 of the document posted for review); then it refers to a
smaller variety of scripts, which may still be a challenge to address
simultaneously.  The scope of script(s) in "Indic" case should be more
clearly defined for the purpose of this study.


3.       Cross script similarity should also be considered, especially in
the case of Cyrillic and Latin, as one could propose TLD strings which are
visually not distinguishable across these scripts, which could be a
significant user issue.


Sarmad Hussain

<<< Chronological Index >>>    <<< Thread Index >>>

Privacy Policy | Terms of Service | Cookies Policy