Registration request for new subtags for Portuguese orthographies

"Martin J. Dürst" duerst at
Thu Mar 26 02:55:46 CET 2015

I'm not totally up to date about the rules for declaring/updating prefix 
fields, but it looks to me like we might get closer to a solution by 
looking at the prefixes.

On 2015/03/26 01:51, João Miguel Neves wrote:
> Just for curiosity, the original proposal (you can also track the
> discussion from there):

The form there (and the one at that's officially 
discussed here this time) has
    Prefix: pt
It seems to me that (part of?) what Michael is arguing is that pt-ao1990 
is very wide and fuzzy, and therefore not very useful.

> On 25/03/2015 16:33, Peter Constable wrote:
>> Michael, consider Shawn's assertion: I need pt-PT-ao1990; I don't need some hypothetical refinements of that.

So if Shawn (and who else?) needs pt-PT-ao1990, wouldn't it be better to 
register this with
    Prefix: pt-PT
Would we need other prefixes at this point? Could we add other prefixes 
if we need them later?

Regards,   Martin.

>> -----Original Message-----
>> From: Ietf-languages [mailto:ietf-languages-bounces at] On Behalf Of Shawn Steele
>> Sent: Wednesday, March 25, 2015 9:07 AM
>> To: Michael Everson; ietf-languages at
>> Subject: RE: Registration request for new subtags for Portuguese orthographies
>>> I was asking you lot to make suggestions. I have objected to the underspecification. I would like you to offer suggestions as to alternatives which would make sense to you. I did not want to get into trouble specifying things that might conflict or be impossible in the kinds of implementations you are working with.
>> For my scenario it is not underspecified.  For Andrew's proposed solution is totally fine for my needs.  I can't invent multiple variant solutions for scenarios I don't have.
>>> So, again, if a user in Portugal wants his variants of ao1990 and a user in Brazil wants his variants of ao1990, how would you propose to respond to those user requirements?
>> I don't, those aren't my user requirements.  They're your hypothetical user's requirements, but I don't have any requests or suggestions or hints of that being a real problem.  Since it's your scenario, I thought you might be able to make suggestions.
>> I need pt-PT-ao1990.  Should your user need pt-BZ-ao1990, that'd be possible.  I presume, however that you mean "three users in Brazil (or Portugal) want different variants of ao1990" since if everyone wanted the same thing we wouldn't be having this conversation.  Again, I don't have that scenario.  I grasp that there are apparently differences I don't quite follow, however:
>> A) There are subtle variants of English & German that nobody ever bothers tagging in practice.  (Indeed the conversation mentioned those earlier).  I'm not qualified to say how "important" further discrimination of Portuguese variants is.
>> B) Should someone need more specificity, then nothing about ao1990 would preclude that.  When such a scenario is identified and proposed, there could be an ao1990-silva (totally hypothetical variation codified by some guy named silva) or ao1990-featureX or ao1990-featureA-featureX (though I can't imagine that level of detail being practical).
>> I can imagine tons of things, but it's not my problem, I don't have the scenarios, I don't have the understanding of the detail you're concerned about, and those details are far too specific for my needs.
>> My scenario is that when a developer updates their app to present a user with a post-reform localization and the user says "gee, the government taught that to my kids in school, but I don't want to play that way", that the developer has the ability to revert the change and provide tags to discriminate for their pro-reform and anti-reform user base.  AFAIK nobody has requested "gee, I love the reform, except for X".
>> I have only two variants to worry about.  For that I need two variant tags (or one I suppose would do).
>> -Shawn

More information about the Ietf-languages mailing list