Request to register private-use variant subtags

[Doug Ewell]

Gordon P. Hemsley wrote:

> Sorry, that was a poor choice of words. What I meant was, you know
> that it's a variant, even if you don't know what it means, and it
> doesn't affect any other part of the language tag. This differs from
> the -x- extension, which could be used for anything, and which makes
> everything after it also be private use.

This is true, and it is one good reason to use "en-Qaaa" for English 
written in Gregg Shorthand instead of "en-x-gregg".

Since the Registry includes subtags for more than 7000 languages, 160 
scripts, and every inhabited region of the Earth, it's usually a good 
bet that a private use -x- subtag (not "x-foo", but "en-x-foo") refers 
to some language variation and not an unencoded language, script, or 
region. That is one reason why support for a private-use variant subtag 
has never been high.

> I know he didn't say that directly. But testing, IMO, is a perfectly
> valid "private use".

Nothing in the Registry or in BCP 47 exists solely to support software 
testing, as this would. It's all about identification of languages and 
language varieties.

> Unicode seems to have asserted by "private agreement" that 'ZZ'
> represents "Unknown or Invalid Territory". [1][2] I see no reason why
> I can't, for example, assert by "private agreement" that 'QZ'
> represents "Testing Region".

I suppose you can, though I'm not sure what the language-identification 
benefit is. You can use existing private-use subtags for your own 
private-use. What we are talking about is introducing new subtags, or 
maybe a new framework, to support testing.

> As far as I can see, I am welcome to privately agree to any meaning
> for any private use subtag.

Any existing one, yes.

I appreciate that you are trying to avoid using invalid tags, and trying 
to avoid collisions (although I think I outlined a reasonable mechanism 
to avoid collisions). I hope you will appreciate that private-use 
subtags have an intended use (tagging languages and variations), and 
that using existing ones for a different purpose (testing software) is 
also allowed, but the standard or its Registry are unlikely to change to 
support that unintended use.

--
Doug Ewell | Thornton, Colorado, USA
http://www.ewellic.org | @DougEwell ­