Re: [Ietf-languages] Fwd: I-D Action: draft-msporny-d-langtag-ext-00.txt

["Doug Ewell" <[e-mail]>]

Manu Sporny wrote:

>> If so, I contend that a plain old language tag should still be a
>> sufficient indicator of that:
>
> Yes, but that was rejected as a solution a long time ago. Some of our
> implementers are running in constrained environments and do not want
> to require a full blown HTML processor to understand language and base
> direction. There are use cases where you can't escape that (such as
> Ruby markup), but for the simple ones... they want a simpler solution
> than "include this gigantic HTML processor in your solution".

Wait, now I'm very lost.

How does the simple language tag "ar" "require a full-blown HTML processor" in a way that the more complex language tag "ar-d-rtl" does not?

>> <span lang="ar">
>> rindfleischetikettierungsüberwachungsaufgabenübertragungsgesetz
>> IS A LONG WORD.</span>
>> (example is HTML, substitute your favorite text format that uses
> language tags)
>
> This is not a solution as it would require a full blown HTML processor
> to understand the variations in an interoperable way

Do you mean that the processor you envision would pick up a BCP 47 language tag, ignore the language part ("ar") altogether, and look only at the direction extension ("-d-rtl")?

If so, this is very much an abuse of the concept of language tagging.

If not, this is more complex than processing "ar", not less.

> (like, for example, what do you do when the tags mismatch, or there is
> a syntax error)?

I don't see how you can avoid solving problems like these, regardless of the approach you adopt.

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