Region subtags and orthographic variants (was: Re: registration requests re Portuguese)

Mark Davis ☕️ mark at macchiato.com
Thu Apr 16 12:52:55 CEST 2015


> pt-ao1990-PT

Syntactically that is impossible for language tags. The ordering must be as
specified in https://tools.ietf.org/html/bcp47#section-2.1.

 langtag       = language
                 ["-" script]
                 ["-" region]
                 *("-" variant)
                 *("-" extension)
                 ["-" privateuse]

That is, it would have to be pt-PT-ao1990. Moreover, there is no chance
*whatsoever* that that would ever change; it is far too embedded in the
world's software.

> need to specify some generally specified departure from 'ideal' (standard)

Of course, internally to a particular domain, you could use your own
protocol for something like BCP47 tags. For external communication, you
would have to translate to and from regular BCP47 language tags at the
domain boundaries, of course.


Mark <https://google.com/+MarkDavis>

*— Il meglio è l’inimico del bene —*

On Thu, Apr 16, 2015 at 1:25 PM, Yury <yury.tarasievich at gmail.com> wrote:

> Doug, spoken like a gentleman. Thank you.
>
> And yes, we are not in a Henry Higgins world here, and every model
> (language tag is, after all, a model of a real world phenomenon) necessary
> loses precision.
>
> And I'm not moving for _wholesale_ throwing out the good old ll_RR model,
> as it still has its uses, and also, because informatics is such a
> traditionalist field.
>
> Now, for your question (I'm not quoting anything else of your elaborate
> text -- for the sake of in-list legibility):
>
> > Does all of this make sense? Any of it?
>
> All of the items 1-5 make sense, I have no quarrel with 1-3 and 5, and
> _my_ contention is only with the proposed practical solution outlined in
> pt.4.
>
> The reason is: ll_RR- and -standard are, in a sense, mutually
> contradicting, _at least linguistically_. That is, those are usus
> (coarsely, real-world departing from ideal) and artifact (ruleset, book,
> paper). You can't particularise first from left to right, than right to
> left, that is confusing.
> (And the thing grates on my nerves, too.)
>
> On the other hand, I agree there might be a need to specify some generally
> specified departure from 'ideal' (standard). Why not do it properly (in the
> sense of building sense hierarchies):
>
> ll-variant-RR, e.g., pt-ao1990-PT --
>
> -- 1) language (the most general formalisation)
> -- 2) state of the pt.1 formalisation
> -- 3) real-world diff from pt.2
>
> Now you have a (fairly) well readable (left to right) scheme, which
> captures all three phenomena and puts them into a proper perspective.
>
> This way, you may particularise even further, in the form
> ll-variant-RR-observation, which would refer to the specific set of
> observation data on the departure from standard.
>
> Which is all in spirit of rfc5646, I believe. I'm only not sure of the
> four part scheme itself; I didn't study it. The respective ABNF spec
> _seems_ to allow it.
>
> (Coming from East European background, I may have a perspective on this
> which many of this list English-world active participants do not share or
> fancy, even.)
> So all of this might be against the lore, of course. Is it?
>
> -Yury
>
> _______________________________________________
> Ietf-languages mailing list
> Ietf-languages at alvestrand.no
> http://www.alvestrand.no/mailman/listinfo/ietf-languages
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.alvestrand.no/pipermail/ietf-languages/attachments/20150416/fcbef123/attachment.html>


More information about the Ietf-languages mailing list