Request to register private-use variant subtags

Doug Ewell doug at
Sat Apr 7 19:54:08 CEST 2012

CE Whitehead <cewcathar at hotmail dot com> wrote:

> I tend to concur with Doug, that rather than registering the whole set
> of gxaaa  - qx999999 variant codes, one would need to redo RFC 5646,
> and include the information about the new reserved codes in that
> document.
> I am unsure as to whether a variant like qxqxqxqx or qzqzqzqz could be
> registered for this purpose.

After discussing Gordon's use case with him privately, I am more 
convinced than ever that no action is needed on the part of either this 
list or LTRU, either to add a subtag (or range of subtags) or to update 
BCP 47.

Gordon wants to unit-test his code against a variant subtag value that 
is not listed in the Registry. (Perhaps strangely, he is regarding the 
existing language, script, and region private-use subtags as "not listed 
in the name database," though of course they are.)

The solution, for unit-testing purposes, is simply to pick a subtag 
value that is not registered, ideally one that is unlikely ever to be 
registered, and optionally to verify as part of his unit test that the 
subtag is not registered. That was Addison's point in saying "A subtag 
like "qxqxqxqx" is a pretty good bet"—not as something that should be 
registered, but on the contrary, as something that will almost certainly 
never be registered.

Does anyone else think any action needs to be taken?

Doug Ewell | Thornton, Colorado, USA | @DougEwell ­ 

More information about the Ietf-languages mailing list