没有不值得去解决的问题,也没有不值得去学习的技术!

Starting Translation Validation for the A Tour of Go Multilingual Project After Selecting Japanese as the Second Language

Figure 1: Multilingual translation project structure in the GitHub repository

作者:

在

After completing the full translation, automated validation, and production deployment verification for the Simplified Chinese version, the A Tour of Go multilingual translation project has entered its next phase: expanding to a second language.

In this phase, I selected Japanese (ja-JP) as the second language. The goal is not simply to add another language directory, but to verify whether the current multilingual architecture can scale continuously.

Moving from a single-language Chinese version to a multilingual translation workflow, the project must solve more than just text translation. It also needs to address:

  • Independent maintenance for different languages;
  • Translation task management;
  • Automated structural validation;
  • Translation quality verification;
  • Continuous synchronization going forward.

Therefore, the Japanese translation phase also serves as a verification of the multilingual architecture.


Selecting Japanese as the Second Language

After completing the Simplified Chinese version, the project already included:

  • Independent locale management;
  • TranslationUnit translation units;
  • Batch translation tasks;
  • Raw response management;
  • Automated validation workflow.

The next phase required selecting a new language to verify that the entire workflow could run independently of the Chinese environment.

The final choice was:

Plaintext
ja-JP

as the second language.

I chose Japanese primarily for the following reasons:

  • Go has a strong adoption base within the Japanese developer community;
  • The Japanese technical documentation ecosystem is relatively mature;
  • It serves as a key validation target for non-Chinese language expansion.

Japanese Translation Tasks Are Underway

I created an independent Japanese translation task directory in the project repository:

Figure 1: Multilingual translation project structure in the GitHub repository
Figure 1: Multilingual translation project structure in the GitHub repository

As shown, the repository now maintains translation tasks organized by language.

Specifically:

Plaintext
data/retranslation-runs/ja-JP/

stores the data generated by the Japanese translation workflow.


ja-JP Translation Batch Management

The Japanese translation uses a batch management approach.

Each translation task has its own independent batch, for example:

Plaintext
chatgpt-ja-JP-001
chatgpt-ja-JP-002
chatgpt-ja-JP-003
chatgpt-ja-JP-004
chatgpt-ja-JP-005

Different batches are used for:

  • Initial translation;
  • Revision fixes;
  • Reprocessing failed pages;
  • New rounds of translation tasks.
Figure 2: ja-JP translation batch directory
Figure 2: ja-JP translation batch directory

This design prevents overwriting historical results while preserving the complete translation process.


First Japanese Page Passes Validation

The Japanese translation workflow does not simply generate text; it requires automated validation.

I have now completed validation for the first batch of pages.

For example:

Plaintext
chatgpt-ja-JP-004

corresponds to the page:

Plaintext
welcome/1

Processing result:

JSON
"validation_passed": 1,
"validation_failed": 0
Figure 3: Japanese page batch validation results
Figure 3: Japanese page batch validation results

Meanwhile, the validation file records:

JSON
"status": "passed"

indicating that the page has passed automated validation.

Figure 4: welcome/1 validation results
Figure 4: welcome/1 validation results

Maintaining a Unified Translation Workflow

The Japanese translation did not introduce a newly designed workflow. Instead, it continues to use the already verified multilingual translation process.

Current workflow:

Plaintext
TranslationUnit
        ↓
ChatGPT full-page translation
        ↓
raw-responses
        ↓
automatic validation
        ↓
Quality Check
        ↓
Final Review

Where:

  • TranslationUnit is the smallest unit for translation and quality checks;
  • batch is the execution unit for GitHub commits, processing, and validation.

The relevant execution workflow is documented in the project files.

Figure 5: Retranslation execution workflow documentation
Figure 5: Retranslation execution workflow documentation

Current Progress

The Japanese version has now completed initial translation validation.

Completed items:

  • ja-JP locale initialization;
  • Japanese translation batch setup;
  • First batch of page translations;
  • Raw response workflow verification;
  • Automated validation verification.

The current focus is not on quickly finishing all pages, but on confirming:

whether the A Tour of Go multilingual translation architecture can reliably support new language expansion.


Next Steps

Going forward, I will continue to push forward with:

  • Batch translation of ja-JP pages;
  • Automated validation;
  • Translation quality checks;
  • Multilingual release workflow verification.

Once the Japanese version passes validation, I will consider expanding further to other languages.

关于作者 · 持续学习与技术实践

我是一名拥有 15 年以上开发经验的技术从业者,自 2013 年起持续运营个人技术博客,记录工作与个人项目中遇到的真实问题,以及解决问题过程中的探索、实践和思考。

除了技术写作,我也在持续开发和维护自己的独立项目,探索新技术和新工具在实际场景中的应用。我始终相信,没有不值得去解决的问题,也没有不值得去学习的技术。

如果这篇文章对你有所帮助,欢迎浏览博客中的其他文章,也欢迎交流技术经验与想法。

关于我  ·  GitHub  ·  邮件联系