Skip to content

Improve sentence-splitting pattern in i18n keys for better SOV + VOS language support #136

Description

@ashb155

Terms

Description

Several i18n keys build a single sentence out of multiple separately-translated string fragments, with a link or resource between the fragments.

This structure assumes every language keeps English's word order (Subject-Verb-Object), with the inline element staying in the same relative position in the sentence. That assumption doesn't for SOV languages like Hindi and Kannada (where the verb moves to the end), as well as VSO/VOS languages (Arabic, Irish, etc.), where the verb needs to move to the front.

Depending on the specific sentence, this produces either:

Fully ungrammatical output, or
Grammatically valid but unnatural phrasing

Example Strings

Image

This issue is related to auditing and consolidating each split sentence into a single translation key with a placeholder for the inline element.

CC @andrewtavis

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

localizationAdd or correct localizations

Type

Projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions