Translating open source project documentation by many translators
summary
The talk describes a pipeline that extracts open-source project documentation from Markdown into a lightweight skeleton and XLIFF translation units, stores matched source→target pairs in a TMX translation memory, and then rebuilds localized Markdown to open a GitHub pull request per repo/branch. The prototype handles a 325,000-package scale, provides fuzzy matching for reuse of slightly changed strings, includes a web editor with Swordfish import, and runs as a working service that automates updates and PR-based publishing.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 2:22Documentation translation problems
- 10:12Standards and solutions
- 16:41CAT tool development
- 22:45How to translate docs
- 29:11English vs localization
- 32:43Q&A
related talks
46mTackling Language Barriers in Open Source Docs: A Case Study of openEuler's...
45mDocs That Scale: Strategies for Sustainable Documentation in Open Source Projects
59mTransforming your Documentation Process
29mCommunity and technology to make document translation work fun and productive- 46mTackling Language Barriers in Open Source Docs: A Case Study of openEuler's...
44mPushing documentation (so it doesn't hurt anymore)
39mOpen Docs, Open Collaboration: Building Better Documentation with Community PRs
33mToward the broader globalization of Open Source