I18n Automation: What need to be automated

Consolidating English content

You can automate the process of collecting and centralizing content updates being worked on in individual branches into a single location in various ways. The complexity of automation depends on how the developer proceeds with front/backend work for a specific feature, and how they go through the process of QA and production release.

Consolidating English content

The process of merging the translated files that have been delivered to a specific branch or repository into another feature branch or release branch (for production release) is a crucial step in achieving automation that satisfies all team members. This process requires a lot of testing and discussion to ensure everyone is satisfied.

The automation of this process is not a one-time setup, and the way and complexity of automation depend on the branching strategy of the manual merge process of delivering translation files to individual branches. Additionally, the entire development team’s process requires additional changes, and each company has different CI/CD (integration & deployment) methods and branching strategies, so there is no golden rule. However, one clear problem that needs to be solved is that developers are having difficulty merging delivered translation files into individual branches, and minimizing this is the ultimate goal of automation.

Pulling EN content

The process of checking out the string repository is to make the targeted English source files recognizable by a String Generation Engine. This is typically done through a remote git checkout, and it is usually the initial step for an external TMS (translation memory system) to access the git repository owned by the company that owns the product.

Generating TMS task with EN content

This is the process of creating translation tasks on a TMS (such as SmartCat, XTM, or Trados), which involves converting the unique format of the source file (e.g. json, js, po, xml, yml, etc.) into a format that is easily recognized by the TMS. It may also include preprocessing to extract added or modified content for improved translation and project management efficiency.

Pulling localized(translated) content

This process involves downloading the completed translation task from TMS to String Generation Engine and performing post-processing to ensure error-free application to the product’s technical requirements. The post-processing can be quite complex depending on the language of delivery or the technical requirements of the receiving product.

Pushing(delivering) localized content

The process of committing the translated files that have completed post-processing back to the string repository, which includes the automation of the entire process of communication for final confirmation needed for delivery, making decisions based on the confirmation, committing, and pushing to the repository.




I am a software engineer who enjoys building things. Please share your interest with me – chrisyoum[at]gmail.com