Publish the Bulgarian glossary with 136 approved terms
Publish the 136 Bulgarian glossary terms marked ready by the translators, with a Bulgarian landing page and interface translation agreed with Claude Opus 5.5. The four terms without publishing approval remain hidden. Other languages' generated content is unchanged.
Translation-lead review: published review page. Reviewers are directed to edit the master sheet directly.
The Glossary Master now has its corrected BG_title header, restored alternative definitions, related terms and source hyperlinks, reference notes, and complete source titles for two entries. Contributor identities and updated links are preserved; Bulgarian credit names now consistently follow the approved Cyrillic spellings in the translators’ Doc. Prose edits are limited to obvious typos. The existing six-person Bulgarian Tenzing record now points to the new glossary. Questions for the translators are collected in work/bulgarian-glossary-20261109/translator-questions.html.
The generator supports a Bulgarian-only refresh and filenames that fit filesystem byte limits. Bulgarian reference notes and mixed-language related-term lists retain the source wording. Broken source links and verified typos are corrected in the master before regeneration. Old URLs are retained for the two corrected titles. Bulgarian headings/navigation mark the English suffix with lang="en", and DefinedTerm metadata declares Bulgarian.
The lead-facing report separates readiness and presentation questions from unverified AI wording suggestions, with resolved items and technical details collapsed. Statistical p formatting, the Schuirmann (1987) citation, and copied Single-blind peer review reference/related-term fields are corrected in the relevant source Doc, master and tracker. FORRT stays abbreviated. Meaning-changing suggestions remain for native-speaker judgment. The detailed automated-review decisions are retained separately in work/bulgarian-glossary-20261109/review-findings.html.
Validation: Hugo production build; all 11 built-HTML checks, including every Bulgarian internal glossary link; 13 focused regression tests plus a rendered italic-p title fixture; sitemap/source date comparison (0 mismatches in 200 sampled pages); desktop/mobile rendering; Cyrillic filtering; reciprocal language links; translated feedback form; all 136 term pages checked for Bulgarian metadata and translator/reviewer credits; source-sheet write readback. Existing languages’ generated content is unchanged. No new build warnings (only the existing Hugo deprecations).
Deployment: after merge, run Data Processing on main with regenerate_glossary=false to refresh the production data artifact with this committed release, then let it trigger deployment. This avoids regenerating unrelated glossary languages.
Created with GPT-6.1 Sol in the Codex harness in T3 Code, with the landing page and interface translation by Claude Opus 5.5.