Extension for modality (was: Re: [Slim] Primary language tags (was "Re: draft-tomkinson comments"))

[Doug Ewell]

Just as a heads-up, the SLIM (Selection of Language for Internet Media)
WG is informally considering an RFC 5646 extension to indicate modality
(spoken, written, signed, possibly others). This would allow indication
of user preference for one modality or the other using BCP 47 language
tags, which is a commonly discussed use case in SLIM.

I suggested the extension approach because IIRC we've already had the
discussion here about adding variants for 'spoken', 'written', etc. and
the consensus was No.

--
Doug Ewell | http://ewellic.org | Thornton, CO 🇺🇸

 
-------- Original Message --------
Subject: Re: [Slim] Primary language tags (was "Re: draft-tomkinson
comments")
From: Gunnar_Hellström <gunnar.hellstrom at omnitor.se>
Date: Sun, November 08, 2015 11:49 pm
To: slim at ietf.org

Den 2015-11-09 kl. 01:32, skrev Martin J. Dürst:
> Sorry forgot one important additional thing:
>
> On 2015/11/09 09:24, Martin J. Dürst wrote:
>> Hello Addison,
>>
>> On 2015/11/04 06:04, Phillips, Addison wrote:
>>
>>> As such, I guess I don't believe that there is much need to introduce
>>> a script subtag merely to say that some set of strings (or collection
>>> of pixels) is written. That much is usually obvious from the resource
>>> itself. Introducing a subtag artificially to convey this information
>>> interferes with tag matching and selection (particularly when the
>>> selection is automated from existing content). And when the
>>> "difference" being identified is already "obvious" to processing, it's
>>> not worth the problems it causes. Use language tags to identify
>>> language variation and not as a proxy for Content-Type ;-).
>>
>> I agree with you with respect to tagging content. But language tags are
>> also about requesting content, and there it can make a huge difference
>> whether something is spoken or written.
>
> There's also the case of subtitled video. It's spoken in one language, 
> and written in another. There are hacks to indicate this with the 
> language tags that we have now (using a script subtag on the written 
> one, and no script tag on the spoken one), but it's quite a stretch.
[GH] And that stretch was what I proposed to use now again for 
distinguishing spoken from written.
The lack of a differentiation between these have kept me from 
considering language tags for relay service selection and invocation 
many years now. Now that we see that there are other applications having

the same problem, let us do something about it.

I can see behind the arguments from the language tagging experts that 
there has been problems to find what you want among already set tags, so

you want to keep the use of extensions low. But I cannot see that the 
difference between spoken and written is a minor detail that should be 
ignored. I think it is more discriminating for the possibility to use 
the content than dialect and script.

Are these motivations convincing enough to start work for a modality 
extension to language tags?

Regards
Gunnar

>
> Regards, Martin.
>
> _______________________________________________
> SLIM mailing list
> SLIM at ietf.org
> https://www.ietf.org/mailman/listinfo/slim

-- 
-----------------------------------------
Gunnar Hellström
Omnitor
gunnar.hellstrom at omnitor.se
+46 708 204 288

_______________________________________________
SLIM mailing list
SLIM at ietf.org
https://www.ietf.org/mailman/listinfo/slim