Click letters in ISO 639-3 and Registry (long)

Summary:

We have an opportunity to make recommendations to ISO 639-3/RA about
changing the spelling of the names of certain African languages to use
the correct click letters and modifier apostrophe, and make the
corresponding changes to the Registry. Please comment if this interests
you.

Details:

Click letters ( ǀ ǁ ǂ ǃ , U+01C0 through U+01C3) were originally
approximated with ASCII fallbacks ( / // /= ! ) in the Ethnologue
language listings, which were the basis for ISO 639-3, which adopted the
fallbacks with few deviations.

Apparently there was a perception within SIL that the proper click
letters were not added to Unicode until 2004 (they have actually been
there since the beginning, 1991), so efforts to update the spellings
were slow in coming.

For many (but not all) language subtags that have Description fields
with click-letter fallbacks, we have added an additional Description
field with the correct spelling. We have retained the fallback spelling
for consistency with 639-3.

The RA did update some of these spellings in 2017, with no announcement,
as it considered these to be display variations and not substantive
changes. These first appeared in data files in December 2017. In the BCP
47 Registry, we are required to match 639-3 names and to make the first
Description field equal to the 639-3 reference name, but are allowed to
modify apostrophes and the like, and have done so a few times.

One of the updated spellings was correct, but most did not use the
correct click letters as they were intended to; instead, they simply
replaced one fallback with another:

ǀ (01C0) was changed from / to | (007C)
ǁ (01C1) was changed from // to || (007C, 007C)
ǂ (01C2) was changed from /= to ‡ (U+2021 DOUBLE DAGGER)
ǃ (01C3) remained as ! (0021)

I have discussed this issue with the ISO 639-3 Registrar, who has taken
it back to the RA, and they agree that if we can provide the actual
correct spellings for the affected languages -- including click letters
and modifier apostrophes -- they will make the necessary changes in the
standard. This does not have to wait for the regular annual cycle.

I am proposing that we recommend these changes to the RA, and that we
subsequently update the Registry (with forms per process) to use ONLY
these correct spellings, with no ASCII fallbacks. We are not obligated
to include an ASCII-only Description field; Section 3.1.5 requires only
that at least one field be in the Latin script, which the click letters
and U+02BC are.

The affected ISO 639-3 code elements and corresponding BCP 47 language
subtags are listed below, each with a proposed disposition. Both active
and withdrawn code elements are involved.

Please review this carefully and make any comments to the list. If we
agree on the "correct" spellings, they will be proposed to the RA in an
appropriate tabular format (no need for Melinda to parse this message
:).


== Currently active subtags ==

Subtag: gku
Description: ǂUngkue
Description: =/Ungkue
ISO 639-3: gku	ǂUngkue

☞	Spelling is already correct in 639-3; no change needed there
☞	Propose removing ASCII fallback spelling in LSR

Subtag: gnk
Description: //Gana
Description: ǁGana
ISO 639-3: gnk	||Gana

☞	639-3 changed from one ASCII fallback (//) to another (||)
☞	We added spelling with ǁ (01C1) in 2015
☞	Propose ǁGana to RA and then use only that in LSR (no ASCII)

Subtag: gwj
Description: /Gwi
Description: ǀGwi
ISO 639-3: gwj	|Gwi

☞	639-3 changed from one ASCII fallback (/) to another (|)
☞	We added spelling with ǀ (01C0) in 2015
☞	Propose ǀGwi to RA and then use only that in LSR (no ASCII)

Subtag: hgm
Description: Hai//om
Description: Haiǁom
ISO 639-3: hgm	Hai||om

☞	Same as gnk; propose Haiǁom (with 01C1) to RA

Subtag: hnh
Description: //Ani
Description: ǁAni
ISO 639-3: hnh	||Ani

☞	Same as gnk; propose ǁAni (with 01C1) to RA

Subtag: huc
Description: =/Hua
Description: ǂHua
ISO 639-3: huc	‡Hua

☞	639-3 changed from one ASCII fallback (=/) to a non-ASCII character
that is still incorrect (‡, U+2021 DOUBLE DAGGER)
☞	We added spelling with ǂ (01C2) in 2015
☞	Propose ǂHua to RA and then use only that in LSR (no ASCII)

Subtag: ktz
Description: Ju/'hoan
Description: Juǀʼhoan
Description: Juǀʼhoansi
ISO 639-3: ktz	Ju|’hoan
ISO 639-3: ktz	Ju|’hoansi

☞	639-3 originally had Ju/'hoan with ASCII fallbacks / and ' (0021,
ASCII quotation mark)
☞	We added spelling with ǀ and ʼ (U+02BC MODIFIER LETTER APOSTROPHE)
in 2015
☞	639-3 added Juǀ'hoansi (with 01C0 and 0027) in January 2017 but
changed that to Ju|’hoansi (with 007C and 2019) in December 2017
☞	Propose Juǀʼhoan and Juǀʼhoansi (with 01C0 and 02BC) to RA for
both names and then use only those in LSR (no ASCII)

Subtag: ngh
Description: N/u
Description: Nǀu
ISO 639-3: ngh	N/u

☞	639-3 has always used just the ASCII fallback (/)
☞	We added spelling with ǀ (01C0) in 2015
☞	Propose Nǀu to RA and then use only that in LSR (no ASCII)

Subtag: nmn
Description: !Xóõ
Description: ǃXóõ
ISO 639-3: nmn	!Xóõ

☞	639-3 has always used just the ASCII fallback (!, 0021)
☞	We added spelling with ǃ (01C3) in 2015
☞	Propose ǃXóõ to RA and then use only that in LSR (no ASCII)

Subtag: vaj
Description: Sekele
Description: Northwestern !Kung
Description: Northwestern ǃKung
Description: Vasekele
ISO 639-3: vaj	Sekele
ISO 639-3: vaj	Northwestern !Kung
ISO 639-3: vaj	Vasekele

☞	Same as nmn, propose Northwestern ǃKung to RA and use that (along
with other names), no ASCII

Subtag: xam
Description: /Xam
Description: ǀXam
ISO 639-3: xam	/Xam

☞	Same as ngh, propose ǀXam to RA and then use only that in LSR (no
ASCII)

Subtag: xeg
Description: //Xegwi
Description: ǁXegwi
ISO 639-3: xeg	//Xegwi

☞	639-3 has always used just the ASCII fallback (//)
☞	We added spelling with ǁ (01C1) in 2015
☞	Propose ǁXegwi to RA and then use only that in LSR (no ASCII)

== Deprecated subtags ==

Subtag: aue
Description: =/Kx'au//'ein
ISO 639-3: aue	=/Kx'au//'ein

☞	639-3 has always used ASCII fallbacks =/ and // and ' (0027)
☞	Withdrawn/deprecated in 2015 and we have never changed spelling
☞	Propose ǂKxʼauǁʼein (01C2, 02BC, 01C1, 02BC) to RA and then use
only that in LSR (no ASCII)

Subtag: gfx
Description: Mangetti Dune !Xung
ISO 639-3: gfx	Mangetti Dune !Xung

☞	639-3 has always used ASCII fallback ! (0021)
☞	Withdrawn/deprecated in 2015 and we have never changed spelling
☞	Propose Mangetti Dune ǃXung (01C3) to RA and then use only that in
LSR (no ASCII)

Subtag: oun
Description: !O!ung
ISO 639-3: oun	!O!ung

☞	639-3 has always used ASCII fallback ! (0021)
☞	Withdrawn/deprecated in 2015 and we have never changed spelling
☞	Propose ǃOǃung (01C3, 01C3) to RA and then use only that in LSR
(no ASCII)


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

_______________________________________________
Ietf-languages mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ietf-languages