Frank Ellermann <nobody at xyzzy dot claranet dot de> wrote:
There is no reasonable chance that ietf-languages will ever make an older (sub)tag the Preferred-Value for a newer one.Not intentionally, but with a valid A -> B and a new subtag C deprecating B you get A -> B -> C, and a simple typo could twist this into A -> B -> A. Admittedly I like graph theory.
Typos in any of the the normative parts of the Registry could cause any number of problems. That's why some of us are watching diligently every time IANA releases a new version, and scrutinizing the registration forms with a magnifying glass before they are submitted to IANA.
IANA is not perfect. Already since the "RFC 3066bis" era began, they have released one with a duplicated record, one with a missing Type field, and one with a multi-line Comments field that was missing the critical leading spaces on the second line. Because we are watching carefully for typos such as these, they were fixed quickly and the number of users that were affected is most likely zero.
Addison Phillips <addison at yahoo dash inc dot com> wrote:
That might not turn out to be true. For example, sometimes countries merge together.
That's a good point. I take back the overly broad statement.Let's consider the case of countries merging. Since Serbia and Montenegro has been such a commonly cited example in LTRU, let's stay with it.
In the next few weeks, or months or something, ISO 3166/MA will decide on new alpha-2 code elements for Serbia and for Montenegro, now that they have agreed to break up. I don't have any inside knowledge on what the code elements will be, but two widely reported possibilities are SP and ME, so let's assume that for now.
Naturally, these code elements will be transformed into RFC 4646 region subtags, but the existing subtag CS will *not* be given a Preferred-Value of either SP or ME. If the two countries -- hypothetically speaking -- should decide to reunite at a later date, the subtags SP and ME would probably be assigned a Preferred-Value of CS, but there would be no circularity problem. I'd be interested if somebody could come up with a plausible scenario that would result in circular P-V assignments.
-- Doug Ewell Fullerton, California, USA http://users.adelphia.net/~dewell/ RFC 4645 * UTN #14 _______________________________________________ Ltru mailing list Ltru at ietf.org https://www1.ietf.org/mailman/listinfo/ltru
Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.