Bug 162331 - Break down the trichotomy between our three current "language groups" (Western, RTL-CTL, CJK)
Summary: Break down the trichotomy between our three current "language groups" (Wester...
Status: UNCONFIRMED
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: LibreOffice (show other bugs)
Version:
(earliest affected)
Inherited From OOo
Hardware: All All
: medium normal
Assignee: Not Assigned
URL:
Whiteboard: QA:needsComment
Keywords:
Depends on: ODF 151215
Blocks: Language-Grouping
  Show dependency treegraph
 
Reported: 2024-08-03 16:39 UTC by Eyal Rozenberg
Modified: 2024-08-18 03:16 UTC (History)
0 users

See Also:
Crash report or crash signature:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Eyal Rozenberg 2024-08-03 16:39:19 UTC
Should written human languages be clearly delimited into three groups: :

* Western languages
* RTL-CTL languages (which are written right-to-left or have complex text layout), and
* Asian languages

? There are multiple loci and contexts in which LibreOffice subscribes to, and enforces, this trichotomy:

* In our font selection UI
* In our documentation regarding languages
* In the ODF spec, where it seems like just about every font property has two extra variants grafted onto it with `-asian` and `-complex` appended to them 
* In our preferences, where extra support for these groups is toggled separately
* Even in our Bugzilla, where many of language meta-bugs are underneath meta-bugs regarding these language groups (until the recent split of bug 43808 into 162322 and 162323).

I argue we should break this trichotomy, generally. We have already acknowledged  this needs to happen to some extent: Bug 151215, "Let me choose different fonts for different languages in the same group". But if you think about it, the same logic applies more widely:

* Users want, and should be able to make, per-language settings - which for now are limited to per-group.
* ODF spec properties and other entities should not hard-code any of this, but rather generalize so that a single languages, all languages or a set of languages can be specified as necessary without replicating names three times ("myproperty, "myproperty-asian", "myproperty-complex" all over the place.
* RTL languages can be non-CTL and vice-versa. In fact, Western languages can be quite CTL'ish when we consider hand writing.
* In fact, some CJK languages, like Japanese and maybe others, are also on occasion written from right to left, i.e. in rl-tb mode (rather than the more popular lr-tb and tb-rl modes).

So the trichotomy is both invalid in principle and in effect.