Macrolanguages (was: Re: BCP 47)

[Phillips, Addison]

>> But these generally do not convey additional utility for the uses to 
>> which language tags are mostly applied and the additional complexity 
>> of macrolanguages is something that generally this group tries to 
>> avoid
> 
> This is NOT the answer that Benson needed.

Actually, I stand behind the first part of that sentence. Even if 'de' were a macrolanguage around Swabian, I would not recommend using it because in most cases it would reduce the usability of the tag in which it appeared.

> 
> This group isn't empowered to avoid the additional complexity of
> macrolanguages. BCP 47 has a mechanism in place to deal with them. The
> mechanism was an uneasy compromise and it does add complexity to the
> tagging model, but when 639-3/RA decides to assign a macrolanguage, that's
> pretty much that; the mechanism kicks in.
> 
> To say that the group tries to avoid the complexity of macrolanguages implies
> that we can somehow override the RA's decision. I don't think we want to
> perpetuate that belief among users.
> 

That's correct. My reply was wrong to imply that. BCP 47 is quite clear about how macrolanguages are (and are not) assigned and such assignments strictly follow ISO 639-3. 

However, I would say that it isn't a good idea to run out and petition the RA to make more macrolanguages. BCP47 has extensive discussion in Section 4.1 about when and how to use macrolanguages: unless your application can derive additional meaning or compatibility by using one with an extended language subtag, it says they are best avoided. 

Addison