Parsing

Parsing is the step that converts a file into a structured representation a system can work with. It reads structured input, such as code or resource files, and interprets its structure so the data can be handled safely and consistently.

In localization, parsing decides what is translatable and what is not. Localization tools use parsers to read formats like JSON, XLIFF, PO, XML, or properties files and distinguish user-facing text from code, placeholders, markup, and metadata. This separation is what allows translators to work on text without touching logic, syntax, or file structure.

Parsing also controls how files are rebuilt after translation. The same parser that extracts strings must be able to place translated content back into the file in the correct location and format. If this step fails, even correct translations can result in broken builds or runtime errors.

👨‍💻 A short example of parsing #️⃣

For example, take a simple JSON file:

{ "title": "Welcome", "retry_limit": 3 }

A parser extracts only "Welcome" for translation and leaves structure and numbers untouched. After translation, it reinserts the text and keeps valid JSON:

{ "title": "Bienvenue", "retry_limit": 3 }

When parsing is inaccurate, problems surface quickly. Code or markup may appear as translatable text, real strings may be skipped, or placeholders may be altered. In many cases, the file becomes invalid after translation and requires developer intervention to fix. This is why localization teams test parsing early and rely on format-aware parsers rather than generic text extraction.