Request: Add retired tag "eml" to the IANA registry

Doug Ewell doug at
Sat Dec 12 18:26:21 CET 2009

I do not see any rationale for adding 150, or any, ISO 639-3 code 
elements that were retired from that standard before RFC 5646 was 
approved.  I strongly oppose the movement to do so.

People who used draft-4645bis as a source for subtags such as 'eml' 
before 5646 was approved, or people who continue to believe ISO 639-3 
code elements *are* subtags and used 'eml' on that basis, have made an 
error, plain and simple.  The purpose of the Registry is not to 
recognize and codify incorrect usage.  We could fill the Registry with 
thousands of wrong entries with pointers to the correct equivalent, if 
we wanted to go that route.

It does not matter whether or not the retired 639-3 code elements are 
guaranteed not to be reused.  They are not suitable candidates for 
language subtags.  When an external standard retires/withdraws/deletes a 
code element that was once a valid BCP 47 subtag, the reason we retain 
the subtag as "Deprecated" is to retain the validity and meaning of 
VALID tags, NOT to maintain a history of the standard.  'eml' was not 
ever part of a valid tag.

Only a very creative reading (in the sense of "creative bookkeeping") of 
RFC 5646, a document which John and many others in this thread 
participated in formulating, could lead to the conclusion that we MUST 
add all (or even some) of the 639-3 code elements that were never valid 
in BCP 47.  Once again, if that were a goal of the LTRU WG, they would 
have been included in RFC 5645; but to the contrary, there is this 
specific wording in Section 2.1 of RFC 5645 (which was also a group 

"Language code elements that were already retired in all of the source 
standards prior to IESG approval of this memo were not listed in these 
files and, consequently, were not considered in this update."

Projects like CLDR may choose to provide an extended list of alias 
subtags, together with their preferred valid equivalents, but this is 
not the domain of the Registry.  People WILL take the presence of 'eml' 
and friends in the Registry as proof they must be valid in BCP 47, no 
matter what annotations we add.

Please do not proceed down this path.

Doug Ewell  |  Thornton, Colorado, USA  |
RFC 5645, 4645, UTN #14  |  ietf-languages @ ­

More information about the Ietf-languages mailing list