LDH-label terminology

Frank Ellermann hmdmhdfmhdjmzdtjmzdtzktdkztdjz at gmail.com
Sun Jul 27 14:26:00 CEST 2008

Andrew Sullivan wrote:

> "the minimum we need for IDNAbis" isn't yet clear to all
> the participants in the discussion anyway.  Or, maybe more
> accurately, there are ideas of what that minimum is, but
> there are several minima in play.

For one minimum in play the syntax is not open for debates,
ordinary 1*63( LDH ) labels not beginning or ending with a
hyphen.  We'd need to agree on a name for it, and some like

Another minimum - there are only two I'm aware of - in play
is 2*63( LDH ) beginning with a letter and not ending with
a hyphen, the "top-label".  Because it begins with a letter
this avoids all known potential issues with IPv4 notations.

So far nobody said that they like minimal length "1" better
than "2", maybe we already agree on this minimum in play...

This so far uncontroversial "top-label" is what's needed to
update RFC 1123 2.1, and get the formal base for IDN TLDs.

> I rather liked Eric Brunner-Williams's taxonomy up-thread.

I'm unsure about using the name "text label" in the context
of the rationale document.  The kind of "text" relevant for
IDNAbis is Unicode with <UTF8-non-ascii> in U-labels.

Eric's "text" in RFC 2929 + 2929bis are "octets interpreted
as case-insensitive ASCII where possible".  If we use this
term we have to add a qualifier, with references to 2929bis
and RFC 4343.

If what you liked also includes "z-label" I'm curious for
what purpose you want a "z-label" concept wrt IDNAbis, i.e.
outside of John's idnabis-alternatives draft.


More information about the Idna-update mailing list