pretranslation

Pretranslation refers to the set of activities performed before the actual translation work begins in a localization or translation project. It aims to prepare source content for translation, ensure consistency, and reduce downstream errors by tackling linguistic, technical, and formatting issues early.

Typical pretranslation tasks include content inventory and analysis, extraction of translatable text from source formats, and the separation of translatable material from non-translatable elements. It also covers text normalization to standardize terminology and punctuation, metadata extraction, and the creation or update of glossaries and termbases to guide translators. In many workflows, a localization kit is produced that includes instructions, style guides, and reference materials. Project setup steps may involve environment configuration, segmentation rules, and alignment of source and target memories.

Often pretranslation includes preparing data for translation memories and machine translation: importing existing translations to populate a memory, or running an MT pre-translation pass that will later be post-edited. Formatting and layout considerations are addressed to preserve placeholders, tags, and metadata in the translated output. The goal is to produce consistent, routinized content that can be translated efficiently while preserving context and formatting.

The scope of pretranslation varies by domain and project. In software and help content, emphasis is on string extraction, UI constraints, and localization of dates and numbers. In documentation, emphasis is on terminology and style. While beneficial, pretranslation must be balanced with risk of locking in outdated terms or over-relying on MT; human review remains essential.