registration requests re Portuguese

[Yury]

On 04/11/2015 12:28 PM, Doug Ewell wrote:
> Yury wrote:
>
>> Isn't any <orthography> just a set of rules, put into the identifiable
>> set of paper books, electronic documents, etc.?
>> People never actually use 100% pure <orthography Mk.X>. The deviations
>> may form their own identifiable ruleset, and so on.
>
> Language tagging can't track differences in usage between individuals.
> That's too fine-grained and would result in tags that are so narrow as
> to be useless for searching. One might as well search by author.

Quite so. There's only so much precision in tagging the text with the 
ruleset ID.

Now, using the 'region' subtag AND ruleset ID 'variant' seems to be 
tentative guess at best. How do you 'place' the text, anyway? By 
internet domain of publishing site? By researching the author's 
background? What if the paper book is printed by trans-nationally 
situated publisher?

I mean the historical use of lang_REGION combination becomes somewhat 
too coarse-grained, and your own example of
pt-ao1990
pt-PT-ao1990
pt-BR-ao1990
pt-CV-ao1990
pt-AO-ao1990
doesn't make sense, somewhat. There are no definable (pinpointable) 
things like pt-AO-ao1990. If there were a ruleset translation or 
harmonisation in AO, then there would be things like pt-ao1990ao or similar.

-Yury