Skip to content

Docs: Upgrade Korean translations from GPT-4o to GPT-5 via Azure OpenAI #201

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Aug 10, 2025

Conversation

skytin1004
Copy link
Collaborator

@skytin1004 skytin1004 commented Aug 10, 2025

Purpose

Upgrade Korean translations using the newly released GPT-5 model from Azure OpenAI to improve translation quality and accuracy compared to the previous GPT-4o translations.

Description

This PR updates all Korean translations in the /translations/ko/ directory using the GPT-5 model available via Azure OpenAI.


Key Improvements Observed

  1. Natural Expressions

    • "우리는 정확성을 위해 노력하지만""정확성을 위해 노력하고 있지만" (progressive form for smoother flow)
    • "재시도""다시 시도" (more conversational)
    • "방법""방법을 사용했습니다" (complete sentence)
  2. Spelling & Terminology Accuracy

    • "루트 디렉토리""루트 디렉터리" (standardized loanword spelling)
    • "아니요""아니오" (Korean orthographic standard)
    • "예제""예시" (better match for context)
  3. Technical Terminology

    • "복제한 경우""클론한 경우" (common developer term)
    • "명령""명령어" (more precise)
    • "글꼴""폰트" (preferred in tech docs)
    • "컴퓨터 비전""Computer Vision" (retain English for industry term)
  4. Clarity & Specificity

    • "활성화""로깅 활성화" (explicit action)
    • "사전 요구 사항""필수 조건" (more natural)
    • "Markdown-Only""Markdown 전용"
  5. Formal Tone

    • "책임을 지지 않습니다""당사는 책임을 지지 않습니다" (explicit subject for clarity)

While GPT-4o was generally quite natural, there were occasional cases where it carried over the structure of English sentences too literally. In contrast, the GPT-5 translations adjust word order and phrasing to better match Korean syntax, resulting in smoother, more readable text. For example, the sentence “우리는 정확성을 위해 노력하지만” (“We strive for accuracy, but…”) became “정확성을 위해 노력하고 있지만,” omitting the unnecessary subject and producing a more natural flow. GPT-5 also adopts terminology commonly used in developer communities, such as replacing “복제” with “클론” and “글꼴” with “폰트,” improving consistency and readability in technical documentation. Nuance handling has also improved — for instance, instead of translating “No” uniformly as “아니요,” it now considers the context of checklists or property fields and uses “없음” (“none”) where appropriate. This context-aware approach leads to more accurate and effective translations overall.


Example Table Improvements

  • Consistent spelling: "아니요""아니오", "아니요""없음" where appropriate
  • More accurate terminology: "글꼴""폰트"
  • Improved column clarity in supported-languages table

Additional Context

  • Model upgrade: Azure OpenAI GPT-4o → GPT-5
  • Changes verified for:
    • Markdown structure preservation
    • Consistency with Co-op Translator style guide
    • Accurate developer terminology
  • Enhances accessibility and usability of Korean documentation for Microsoft and open-source communities

Type of Change

  • Bugfix
  • Feature
  • Code style update
  • Refactoring (no functional/API changes)
  • Documentation content changes

Checklist

  • Thoroughly tested all updated translations
  • Verified no formatting or structural regressions
  • Followed repository’s style and terminology conventions
  • Updated documentation accordingly

@github-actions github-actions bot added the documentation Improvements or additions to documentation label Aug 10, 2025
@skytin1004 skytin1004 changed the title Upgrade Korean translations from GPT-4o to GPT-5 via Azure OpenAI Docs: Upgrade Korean translations from GPT-4o to GPT-5 via Azure OpenAI Aug 10, 2025
@skytin1004 skytin1004 merged commit 048236d into Azure:main Aug 10, 2025
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
documentation Improvements or additions to documentation
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant