Harald Alvestrand harald at alvestrand.no
Wed Mar 22 16:18:51 CET 2006

Doug Ewell wrote:
> Not to beat this into the ground, but the word MODIFY in the title of 
> this message -- as originally sent by me and forwarded by Michael to 
> IANA -- is IANA's clue that this request is for the *modification* of 
> an existing subtag, not the *addition* of a new subtag.  (That is also 
> why the Added date for this record was not 2006-03-08 as it was for 
> all the other requests.)
> The result of treating this modification request as an addition is 
> that there are now two records in the registry for language subtag 
> "fy". This changes the registry from being out-of-date to being 
> invalid.  It needs to be corrected by either:
> 1.  removing the first (older) "fy" record, or
> 2.  removing the second (new) "fy" record and changing the description 
> of the first (older) record from "Frisian" to "Western Frisian."
> Mark's idea of giving IANA a new, ready-to-eat file that they simply 
> need to host on their site starts to look better and better.
Note - this is not related to experience with language tags, but to 
experience with other registries....

If you instil knowledge of what's "right" in a body with institutional 
memory, such as I hope IANA has, it gets passed on.

If you instil that same knowledge in an expert who has no organizational 
relationship with the next expert, the knowledge transfer can get very 
spotty - th next expert can easily forget about the details of handling 
the registry; he's usually a technology expert, not a procedural expert.

This is most often a problem when there are gaps of years between 
registrations, so that an expert can have forgotten that he's the 
designated expert before the next request comes in - but the principle 
is general.

Suggestion..... why not write a Perl script or other tool to verify the 
"goodness" of the registry, and offer that to IANA for their use?


More information about the Ietf-languages mailing list