RE: [idn] First report from IDN nameprep design team

[Paul Hoffman / IMC <phoffman@imc.org>]

At 8:10 AM +0200 12/7/00, Jonathan Rosenne wrote:
>I repeat my request to ignore certain Hebrew characters, namely points and
>accents, i.e. remove them during nameprep.
>
>I think some Arabic experts have indicated that this should be 
>applied also for
>Arabic.

If this WG adopts the proposals from the design team, we would have a 
much cleaner way to do what you suggest. These characters can simply 
be mapped to nothing in the mapping step.

However, we have to be careful not to just map out the vowels and say 
we are done. If we did so, and someone uses a character like U+FB2A 
(a Hebrew presentation form) or like U+FC5D (an Arabic presentation 
form), the points would make it into the final name part. Thus, if we 
choose to ignore some some characters, we have to also carefully find 
all characters that whose normalizations would include the ignored 
characters and add maps for them in the mapping step. Fortunately, 
this can be done programmatically.

--Paul Hoffman, Director
--Internet Mail Consortium