frr, fy, ngo, tt (was: REQUEST for registration of variant subtag 'grabar')

[Frank Ellermann]

Michael Everson wrote:

> We have precedent, too. I rejected the Frisian suppress-Latn
> because it was done to test the system.

Oops, I almost forgot that waiting for the 4646 publication,
thanks for the reminder.  The test period is certainly over,
and there is now a precedent where that detail was solved as
expected:

Type: language
Subtag: nqo
Description: N’Ko
Added: 2006-06-05
Suppress-Script: Nkoo

In <http://permalink.gmane.org/gmane.ietf.languages/2564> you
wrote (2006-04-03):  "Subsequently people began to work to do
a more comprehensive job with the Suppress-Script material with
a view to making a more comprehensive set of applications. My
judgement is that frr can wait for that task to be completed."

Apparently nothing happened with that task in the last six
months, and the last CLDR 1.4 list also has no script for frr:

| <tr><td class='target'>[frr]	Northern Frisian</td></tr>

No ngo yet, it's probably older.  That list has fy as Latn:

| <tr><td class='source' colspan='2'>[fy]	Western Frisian</td></tr>
| <tr><td>&nbsp;</td><td class='target'>[Latn]	Latin</td></tr>

It also offer tt as Cyrl, simply copying it wholesale in a
bulk update of the registry for 639-3 would be questionable.

For languages with one dominant script it's better if that's
documented in the registry, copied as needed by the CLDR folks,
not the other way around:

For the initial registry "copy CLDR" was an exception, with an
IETF Last Call, documented in RFC 4645.

Frank