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

Re: Character Variant Deployment at VeriSign



Edmon Chung <edmon@xxxxxxxxxx> wrote:

> Just as a hypothetical example:
>
> a registry may have the following policies:
>
> if domain=EPA.tld + lang=en --> varset={none}
>
> if domain=EPA.tld + lang=el --> varset={<Eta><Pi><Alpha>.tld;
> <Ie><Er><A>.tld}
>
> if domain=EPA.tld + lang=ru --> varset={<Ie><Er><A>.tld}
>
> In the above example, the variant set for a given domain "EPA.tld"
> could be different if the language tag is different.
>
> Without the language tag, how can the variant set be determined?

Okay, I'm starting to see something.  I think anyone who registered EPA
and anyone who registered Eta Pi Alpha would be likely to have a dispute
regardless of the intended languages, and some users would be likely to
be confused regardless of the languages intended by the registrants.  So
I don't think language tags are useful for deciding what registrations
get blocked.

However, when the registry suggests variants to actually add to the DNS
zone (unlike most variants, which would be blocked but would not appear
in the zone), or when the registry sets prices for adding particular
variants to the zone, I can see how it would make sense to use a
language tag.

So I suggest a model where language tags might be relevant to the
internal affairs of a single registration (like pricing and the domain
management user interface), but would be irrelevant for interactions
between registrations (like the construction of the full set of blocked
variants).  What do you think?

AMC