From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Wed Nov 23 2005 - 05:47:47 CST
From: "Kenneth Whistler" <kenw@sybase.com>
> It is another question to consider what the best thing would
> be for Yiddish in IDN. If you are asking ME, I would suggest
> that simply spelling out the digraphs as sequences of vav
> and/or yod and disallowing the digraphs in IDN would quite
> clearly cut down on the possibilities for spoofing in
> Yiddish domain names. It might, however, marginally raise the level
> of user confusion in attempting to input Yiddish domain names.
> And since user confusion is much of what IDN is attempting
> to avoid, I suspect there is simply a tradeoff here with
> no optimal answer.
This could be addressed differently: the IDN could include an extra set of
decompositions which are not in Unicode. Ofcouse the IDN authors would seek
for advice at Unicode to see if those extra decompisitions affect the
linguistic. But if there's no evidence that this may be detrimental to
Yiddish (or possibly other languages using those ligatures), what would be
wrong if the NamePrep algorithm included extra decompositions?
Is it up to Unicode to fix that? May be Unicode could simply publish its
list of confusive characters, and let third-party applications (like the
IDN) which of them to apply in their algorithms. (Unicode already documents
several folding algorithms that are not normalizations, and that all are
optional, whever to use them or not is left to the users)
This won't affect the statiblity of Unicode itself, but is the IDN
specification *stable* for now? I know it has several specifications, but
many revisions as well. Is there a registry which accepts those Yiddish
ligatures for now?
This archive was generated by hypermail 2.1.5 : Wed Nov 23 2005 - 09:05:13 CST