Date: 2026-08-06 20:34 Summary: Current pages aim for search-habit localization rewrites, not pure translation; quality is mixed and ar/ja/ru lack verified native review.
[question] 现在页面的语言是翻译的，还是根据本地化用户习惯和搜索习惯写的？
[try to solve] Target: rewrite for local search habits, not English-template literal translation. Delivery is mixed, and live native human review is not yet verified.
Rule intent
tool-i18n-localization.mdc requires: locale brief (query / title direction) → independent rewrite from the brief → ≥3 rounds; forbid Translate to {lang} and scripted isomorphic dumps.
Live evidence (leans localization)
- Titles use local primary terms: es/pt/fr use IMC; ru uses the Russian BMI acronym; zh uses public-IP and BMI phrasing natural to Chinese search; ja uses natural “how to get body-mass index” / “when to use this” headings
- Some FAQ topic order already differs (e.g. marginal revenue: es/id/de vs en)
- Stock-tool
03files mostly marki18n-done; relative to early.bakliteral titles, search-oriented rewrites landed
Still translation-skeleton-like
- Module structure stays largely parallel across ten locales (How / Formula / FAQ topic sets often match)
- Many FAQs are the same question set with different wording (e.g. percentage-change)
- ar/ja/ru “pending native review” was cleared by user confirmation — not completed live native QA
One-liners
| Framing | Answer |
|---|---|
| Pure machine translation out? | Per current process and spot-checks, not the delivery target |
| Native original for search habits? | Intent and title layer are close; long copy/FAQ still show isomorphic traces; no verified live native review |
More accurate label: search-oriented AI rewrite (localization-oriented rewrite), not professional localization final copy.
[actions]
- Spot-checked
src/site/i18n/*title/FAQ/usecases headings; comparedwork-tasks/*/03-locale-briefs.mdand localization rules