[Ietf-languages] Re: Braille types

[Hugh Paterson III <[e-mail]>]

Thank you for these clarifications. I will make the proposal. But this week
I am traveling so next week at the earliest. If there are additional
considerations then i am all ‘learners ears’.


All the best,
-Hugh

Sent from my iPhone

On Mon, Jul 20, 2026 at 1:05 AM Mark Davis Ⓤ <[e-mail]> wrote:

> There is no requirement that a variant specify a prefix. It would be okay
> to have a variant bgroupe1 with a description that it is used in language
> tags that have a Brai script tag.
>
>
> BCP 47 specifies that a language tag containing a variant *should* match
> at least one of its registered prefixes. It deliberately uses *"SHOULD"*
> (as defined in RFC 2119), rather than *"MUST"* or *"SHALL"*.
>
>    - ​A "SHOULD" condition means that while matching the prefix is
>    recommended for best practices and broad compatibility, omitting or
>    altering it *does not break formal validity*.
>
>
> On Sat, Jul 18, 2026, 07:31 Hugh Paterson III <[e-mail]>
> wrote:
>
>> Thank you Doug for these comments.
>>
>> > The definition of the Prefix field cannot be changed except by
>> revision of RFC 5646. That’s not to say that a revision has never been
>> considered, or could never happen, but the IETF process does make such an
>> effort difficult and unappetizing.
>>
>> I understand the unappetizing nature of revisions.
>> 1. Is there a list of "things to modify" in the next revision process
>> whenever that happens?
>> 2. Is there an amendment process; or erata process which might "fix" this
>> without major revision?
>>
>> - Hugh
>>
>> On Thu, Jul 16, 2026 at 10:17 PM Doug Ewell <[e-mail]> wrote:
>>
>>> Hugh Paterson III wrote:
>>>
>>> > The issue with the -fonipa variant tag is that it doesn't specify
>>> > which version of IPA is used in the text.
>>>
>>> In the 20 years since 'fonipa' was registered, I don’t think I’ve heard
>>> anyone complain that it was unusable due to not being specific enough.
>>>
>>> > Why is the domain of a subtag the language entity identifier instead
>>> > of optionally being a modifier of the script entity identifier? If
>>> > subtags can be scoped so that they can only be used in conjunction
>>> > with certain language identifiers, why can't they also be scoped so
>>> > that they can only be used with certain script identifiers? With this
>>> > line of reasoning I am not talking about "across languages" I am
>>> > talking about across the script "Brai" and wherever/whenever "Brai" is
>>> > used.
>>>
>>> The Prefix field is defined in Section 3.1.8 as containing “a valid
>>> language tag” and not an extended language range (RFC 4647, Section 2.2)
>>> such as "*-Brai". One might make a strong argument that this was
>>> shortsighted, that a variant subtag might bind much more strongly to script
>>> than to language and a Prefix should be able to reflect that.
>>>
>>> The definition of the Prefix field cannot be changed except by revision
>>> of RFC 5646. That’s not to say that a revision has never been considered,
>>> or could never happen, but the IETF process does make such an effort
>>> difficult and unappetizing.
>>>
>>> > What is common across language contexts in which Braille is used as a
>>> > script, is that they have (usually---at least) two different
>>> > transcription systems upon which digital text decisions are made.
>>> > There exists a contracted form and an uncontracted form. It seems more
>>> > efficient to create two or three subtags which are scoped to a script
>>> > than to create 200-300 subtags which are scoped to languages.
>>>
>>> OK, I’ll take a swing at that pitch.
>>>
>>> Yes, you're right and I was wrong, to an extent — if there is a
>>> practical need to provide discrete BCP 47 tagging for uncontracted and
>>> contracted varieties of Braille as used for each of 60 to 100 or more
>>> languages, then it would be better to provide one subtag for “uncontracted”
>>> and another for “contracted,” and possibly (though less convincingly) a
>>> third for “personal, idiosyncratic, seldom-published shorthand,” than to
>>> multiply X by Y and propose 200 to 300 variant subtags most of which will
>>> never be used. (Note that there are only 139 variant subtags in existence
>>> today.)
>>>
>>> Some sort of subtag values that convey “uncontracted” versus
>>> “contracted” would be preferable, in my mind, to “level1” and “level2,”
>>> which imply those are the commonly used names. If many other
>>> language-specific implementations of Braille use the terms “Grade 1” and
>>> “Grade 2” as English Braille does, then those might be suitable.
>>>
>>> > When thinking about subtags, it may be that 200-300 subtags are
>>> > required. But it will likely be more than that because, as is the case
>>> > with French Braille, there are different "orthography" variants for
>>> > different time periods. I haven't personally investigated the
>>> > situation beyond French, but suspect that there are "evolutionary
>>> > stages" in other languages as well.
>>>
>>> Of course there are, but whether there is demand for BCP 47 tagging of
>>> those historical stages is another matter. The Registry is not a research
>>> work documenting all known language varieties. It exists for practical
>>> tagging purposes.
>>>
>>> “Baseball” was originally spelled “base ball,” and later “base-ball.”
>>> Words like “today,” “tonight,” and “tomorrow” were spelled with hyphens as
>>> recently as the 1940s. Not all orthographic varieties throughout time make
>>> the case for being taggable differences.
>>>
>>> --
>>> Doug Ewell, CC, ALB | Lakewood, CO, US | ewellic.org
>>>
>>>
>>> _______________________________________________
>> Ietf-languages mailing list -- [e-mail]
>> To unsubscribe send an email to [e-mail]
>>
>