Re: [Ietf-languages] d-langtag-ext Use Case (was: Re: Fwd: I-D Action: draft-msporny-d-langtag-ext-00.txt)

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

Manu Sporny wrote:

> We requested a review of the Verifiable Credentials specification (via
> [e-mail]) in 2018 and again in Feb 2019.

I am sorry that your timely requests for review went unanswered. Sometimes the wheels turn slowly through no fault of the requester.

> I'll refrain from being flippant and not say "That you're just hearing
> about this is 'not my problem'... how was I to know where you hang out
> on the Internet and that you had an opinion about this? Why aren't you
> also hanging out in public-i18n-core?". :) -- but, I don't think doing
> stuff like that will get us to better i18n, which is what I'd like us
> to focus on. :)

I apologized before for that unfortunate expression, and I apologize for it again now. It was a bad choice of response; it was neither productive nor indicative of willingness to help. I'm sorry.

Many of us, and I'm sure you as well, have heard requests over the years of this form: "We have use case A, and we propose to solve it with mechanism B." Use case A may be completely genuine and cause real-world problems, but it does not necessarily follow that mechanism B is a good or suitable solution. Repeating the arguments for use case A and emphasizing the timeliness factor does not change mechanism B into a good solution.

We are trying to explain alternative solutions that will satisfy use case A.

It is certainly not your fault or anyone else's that we did not include something in Section 3.7 of BCP 47 about encouraging authors of extension RFCs to contact [e-mail] for feedback and review. BCP 47 does mention the ietf-languages list in more than a dozen places, so for those who have read the document, it might seem like an appropriate place to send a request for such feedback. We are not just all over the Internet.

> This is an attempt to see if the extension would work for the set of
> use cases we're grappling with and then provide some guidance in our
> specification.

Here is the question that keeps popping into my head (also when following the thread on rdf-dir-literal):

How does a language tag that just indicates the language, or language plus script, NOT solve the need to indicate base directionality?

Mark gave an example of a short paragraph that doesn’t begin with the "right" directionality, so an attempt to auto-detect would guess wrong. But if you tagged that paragraph as "ar", shouldn't the base directionality be that of Arabic — that is, RTL?

> Here's the use case if you (or anyone else) is still interested in helping:

It does seem to me that on the GitHub thread, the use case is being framed in terms of the proposed solution. Here is an example:

"you cannot necessarily get the direction information just from the language."

"Sorry, I meant that this is already true for many languages, but not all. For those languages where it is possible, that information comes from the language tag (e.g., "en" means the direction is "ltr"). That indicates to me that a sensible solution to this problem is to add the direction information to the language tags that do not imply it already (via an -d- language tag extension)."

My question is, for what combinations of language and script is it *not* possible to figure out the directionality? 

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