Re: [ga] whois: issues with uniformity

["J-F C. (Jefsey) Morfin" <jefsey@club-internet.fr>]

At 11:11 23/12/02, Thomas Roessler wrote:
>For an internationally useful WHOIS service, you'd most likely like
>to be able to accomodate /both/ versions in the database.
>
>This isn't about Unicode-enabling a database: This is about the
>human brain in general /not/ being Unicode-enabled, and about
>accomodating existing real-life band-aids for that.

Correct. Your point is extremely well done. The first time I see it risen?
This is geting real: what IETF has difficulty with, being "engineers".

Another problem is "babelling", ie using the ascii transcoding not just as 
a transcoding but for its real ascii value, and registring the Unicode in 
consequence. Something Stephane avoided to commit about at IETF (to say if 
I am right or wrong about this permitting non prosecutable cybersquatting :-)

This is simple: you may register an Unicode sequence because it has a 
meaning in a language. You may also register one because its transcoding in 
ascii has a meaning in ascii, since on every system not supporting the 
concerned character set will print it as ascii.

Exemple:
u+0067 u+0077 u+1E21 u+1E1F.com has probably no real meaning in any 
language. But it prints as "iesg--gw-bush.com" what makes it fun.

u+0063 u+006F u+0B69 u+0B69 u+0063 u+0061.com also has no meaning and 
difficult to print in Unicode, so it will print in most of the cases as 
"iesg--coca-cola.com" what makes it a non-UDRP candidate (it has never been 
registered in ascii :-) a true cybersquatter paradise.

The best is u+3A17.com which will also print as "iesg--ibm.com".
no prosecution: kust a check. probably for every TLD fully supporting IDNs.

jfc