[Ltru] Re: Test suite for language tags?

John Cowan wrote:

> My point is that fr-Lat is not a *grandfathered* tag.
> Nothing can ever be a grandfathered tag except the 34
> existing examples.  The current ABNF for a grandfathered
> tag, 1*3ALPHA 1*2("-" (2*8alphanum)), therefore recognizes
> too much.

Okay, now I got it.  We could create more elaborated rules
with less mismatches, something in the direction of "if it
doesn't start with art / cel / en / i / no / sgn / zh it's
certainly not grandfathered".  And that won't match fr-Lat.

Or we split it into i / sgn / zh, adding the rest as direct
enumeration (1*art, 1*cel, 3*en, 2*no = 5 exceptions).  But
then we'd be in a situation if boont / gaulish / scouse are
registered as ordinary variants.

We could also decree that boont / gaulish / lobjan / scouse
are dummy + deprecated variants, this would move 4 of the 34
to redundant.  Maybe also guoyu / hakka / xiang for 7 of 34.

We had reasons why we didn't do this for 3066bis, but if you
have a better plan that could be changed in 3066ter.  There's
no chance to get rid of en-GB-oed / no-bok / no-nyn.

> Listing them would do no harm.

Listed they are, you propose to copy the list into the ABNF.
For en / no / sgn / zh I don't see a perfect syntax solution.

Frank

_______________________________________________
Ltru mailing list
Ltru@ietf.org
https://www1.ietf.org/mailman/listinfo/ltru