Retired 639-3 codes

Michael(tm) Smith mike at
Mon Dec 14 07:14:49 CET 2009

Doug Ewell <doug at>, 2009-12-13 22:11 -0700:

> Michael(tm) Smith <mike at w3 dot org> wrote:
> >> aex      Amerax eng
> >> [...]
> >> yib      Yinglish       eng
> >
> > Well, those are certainly interesting. I wonder how many other cases 
> > that list might have in which the preferred value cited is itself an 
> > invalid tag.
> 'eng' is a perfectly valid code element in ISO 639-3.  It is not a valid 
> subtag in BCP 47, but John gave us a list of ISO 639-3 code elements, 
> not BCP 47 subtags.

OK, understood. I didn't know about that difference. I'm certainly
starting to really appreciate the complexities around all this...


> > Was "eng" ever valid? Is it in a different class of 
> > retirement/deprecation than the "eml" case?
> Again, be careful not to confuse ISO 639-3 code elements with BCP 47 
> subtags.  'eng' is and has always been a valid ISO 639-3 code element. 
> It is not and never will be a valid BCP 47 subtag, because 'en' --  
> derived from the ISO 639-1 two-letter code element -- is used instead 
> (for reasons that would be both very difficult and very ill-advised to 
> try to change).  There is no retirement or deprecation involved here.

OK, I see.

I also see now that my thought about trying to use the ISO 639-3
code retirement mappings list[1] to provide suggestions for the
retired codes wouldn't work on its own anyway -- because I'd then
also need to have some other data source to be able to map a
suggested-replacement code element such as "eng" to its equivalent
BCP 47 subtag.


So that implementing the (mis)feature I originally had in mind for
the case of reporting on the retired codes is probably just not
worth the trouble, and quite possibly not even a good idea even
were it more easily doable.


Michael(tm) Smith

More information about the Ietf-languages mailing list