proposed ISO standard for language variations

[Doug Ewell]

Peter Constable wrote:

> Follow-up on this topic:
>
> L2/16-242
> Ballot results for TC37/SC2 New Work Item Proposal: Identification and
> description of language varieties

Yeah, saw that.

It's still not clear to me how supporters intend to work this into BCP 
47, or to what extent they understand the language identification 
options already available. Comments about benefits to the linguistics 
community seem sensible -- those folks can always use more detail -- but 
some NBs wrote that this granularity is critically important to tagging 
of software and IT resources. I would not have guessed that, especially 
for varieties like stuttering or "motherese" or "individual speaker."

There is a perception that BCP 47 encourages (not just "permits") 
excessively long and detailed language tags. It's known that some 
specifications that reference BCP 47, such as other RFCs, feel it 
necessary to add extra "health warnings" about keeping tags short, even 
restricting them to primary language subtags only, ignoring Section 4.1 
of RFC 5646 which already advises against using unnecessary subtags. The 
notion of applying this NWIP to BCP 47 seems to go in the opposite 
direction.

If this is issued as a standard, naturally there would have to be some 
sort of registry for values like "business speak" or "Two-word-stage," 
and some way of handling the really open-ended values like "John Doe." 
Shoehorning the speaker's name into extension subtags isn't as 
straightforward as it might seem: what if the speaker's last name is 
"Constable" and thus longer than 8 letters? What if there's more than 
one Peter Constable out there?

We would always have the option of rechartering the LTRU WG to create a 
new BCP 47 spec (replacing 5646) to support ISO XXX inherently, creating 
a new extension spec per Section 3.7, or simply ignoring ISO XXX as out 
of scope for BCP 47.

--
Doug Ewell | Thornton, CO, US | ewellic.org