pseudo-localization variants

[Doug Ewell]

I fear that once again, the horrors and dangers of private-use subtags 
have been grossly overstated. Section 4.6 of RFC 5646 does talk about a 
"private agreement" that can exist between consenting adults (as does 
the Unicode Standard). Private-use is neither forbidden nor useless 
outside of one's own home, as long as there is an agreement stating what 
the tags mean.

There is a big difference between invalid tags like "qps-ploc" and 
"qps-plocm", and valid private-use tags like "qps-x-ploc" and 
"qps-x-plocm", and I'm disappointed that Microsoft would be using the 
former, six years after the publication of RFC 4646. A valid private-use 
tag can at least be parsed properly, and the private-use parts 
interpreted or ignored as appropriate.

Note that I'm not judging whether subtags for pseudo-localized text 
should or shouldn't be encoded. I can't tell if this situation is more 
like "computer languages" or "programming languages" which have been 
ruled out since RFC 1766, or like "i-default", enshrined as a taggable 
entity almost 15 years ago. But if the decision is made not to encode, 
please, everybody, use valid private-use tags; don't just throw up your 
hands and make up valid-looking tags that aren't valid.

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