[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Comparison of hoffman-idn-reg and jseng-idn-admin

>>At 7:01 PM +0430 4/3/03, Roozbeh Pournader wrote:
>>>Can't a registry just have a complicated script that after one asks
for a
>>>label to register tells: Sorry, you can't register this label,
because it
>>>"contains the word sex" or "uses FINAL KAF in a non-final position"?
>>This is an important point. Registries do not have to blindly follow
>>any rules, such as the ones in this document or the JET document.
>>They can look at the output bundle and make local policy decisions.
>>For characters that are position-dependent in words, a registry will
>>have to have non-automatic rules. For example, assume that the letter
>>"z" could not be at the end of an English word. "baz.us" should be
>>not allowed, but the .us registry couldn't just check for "is 'z' at
>>the end of the word" because that would allow someone to register
>>"baz123.us". Or assume that "y" can only be at the end of a word. The
>>registry should not have a rule that disallows any name that has "y"
>>in the middle because someone might correctly want to register

I agree with Paul. Unless you ban in your rules *all* registrations of
two consecutive words in a label (i.e., without any non-letter
separation between them) then an automatic block of only the final forms
will lead to "discriminatory" blockings of forms that are legit (in
Paul's example "tellme.us" will go through, but "floycorp.us" will not,
unless you don't allow these forms and allow only "tell-me" or "tell_me"

This discussion also seems to assume that there is a requirement in the
rules that labels must have *some* meaning. 

Having the above rule as binding, and utilizing it automatically, seems
as a too strong limitation on the shape of the name space. On the other
hand, unless the registry has no limitation on the shape of a label,
this particular "corner" seems to indicate that human supervision and
discretion in these languages is necessary.