Doug Ewell doug at
Fri Dec 30 04:09:07 CET 2016

Shawn Steele wrote:

> The -c- seems a bit clumsy to me, especially when es-c-en and en-c-es
> are supposed to be equivalent. BCP47 and the like sort of indicates
> that the -c- stuff could be ignored. So some systems would see those
> as "English" or "Spanish" and lose the minor detail of the combined
> form.

The fallback problem is the same for variants like 'spanglis' and for
extensions, unless you take Kent's idea of starting the tag with "mul"
and then adding the variant or extension. Then at least the fallback is
"mul", which is true, but people don't like "mul" under any

The only solution that addresses this issue is to declare the hybrid to
be a language, and assign it a 5-to-8-letter primary language subtag.
But that has problems as well: (a) list members are afraid to register
such a thing, (b) the hybrid might not really be a language, and (c)
it's still a one-off solution, not suitable if we think there are many
such hybrids.
Doug Ewell | Thornton, CO, US |

More information about the Ietf-languages mailing list