[Ltru] Re: Remove extlang from ABNF?

Felix Sasaki wrote:

> Maybe you are not aware of the situation that language tags are nearly 
> the only IETF spec where W3C uses a BCP reference, or before that an 
> "RFC XXX or its successor" reference - in difference to IRI, btw .

Yes, some including me proposed standards track for 4646, but the
rough consensus was BCP :-(  You use 2119 (BCP 14) quite often in W3C
spec.s, but that's no moving target.  Unlike RFC 4646, were "moving
target" was predictable at least for extlang:

| reserved for future standardization, anticipating work that is
| currently under way on ISO 639.
[...]
| Their syntax is described here so that implementations can be
| compatible with any future revision of this document that does
| provide for their registration.
[...]
| "zh-gan" (registered under RFC 3066) might become a valid non-
| grandfathered (that is, redundant) tag in which the subtag 
| 'gan' might represent the Chinese dialect 'Gan'.

Note "might", it's not "shall".  We weren't sure, and the reasons
to drop "extlang" are rather complex (e.g. 639-3 is more volatile
than expected with its macrolanguages at the moment, not good for
a "stable forver" registry).

I still don't see the harm if an implementation based on RFC 4646
identifies a bogus (non-existing) tag as "well-formed", where a 
4646bis implementation would see "not well-formed" with the clean
ABNF proposal.  And both editors would apparently prefer "clean".

> You might say you don't care about this status

And I did, see above, 4646bis could be a nice "draft standard"
soon.  Dropping unused features in a DS is perfectly okay.  It's
also okay in a BCP, but apparently you tjink it's not.

> reading Ira's mail, it seems other people are in a similar 
> situation.

I'm not sure in what situation she would be with a simpler ABNF,
she used to be a "KISS" fan.  I've no idea what a printer could
do with a "well-formed" 4646 extlang tag when no extlangs exist.

> Regarding the change of "irregular", this seems to me much 
> easier to sell to people than deleting the extlang production.

In -10 it's still "extlang = [...] permanently reserved", in -11
it might be "unused = [...] permanently unused".  Only to keep
constructs well-formed that permanently won't exist.  Some folks
were less timid while introducing IPv6 literals in URLs, and W3C
documents use all kinds of mixtures, pure 2396, 2396 as amended
by ????, and the real thing, STD 66 (3986).  

After all "RFC" still stands for "request for comment", even BCPs
change (RFC 3066 was BCP 47 before, RFC 1766 was a proper PS).

 Frank

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