Terminology and glossaries

Terminology management is the work of deciding how key words are written and translated, and then making sure everyone follows those decisions. A termbase holds each concept with its approved term in every language, plus notes such as definitions, forbidden alternatives, part of speech and where the term is used. When translators, writers and machine translation all draw on the same termbase, a product feature has one name in each language instead of four.

The tools here cover different points in that process. Trados MultiTerm is a dedicated terminology system that builds termbases for Trados projects, and a good termbase system adds approval workflows so new terms are requested, reviewed and released. Coreon organizes terms as a map of related concepts, which helps in technical fields with deep vocabularies. Acrolinx and Congree check source writing against terminology and style rules before it reaches translation, which stops inconsistent terms at the start. Sketch Engine is a corpus tool that helps terminologists find candidate terms in large collections of real text.

Start small. A list of a few hundred terms that matter most, such as product names, feature names, terms that must stay in English and words customers often see, does more good than an attempt to capture every word. Decide who can add and approve terms in each language, usually a mix of a central owner and reviewers in each market.

Then connect the termbase to the tools that produce text. Most translation management systems read termbases in TBX format and highlight terms for translators. Some machine translation engines accept glossaries so the approved term is used automatically. Writing assistants can flag the wrong term while a source author is still drafting. A termbase that lives only in a spreadsheet tends to drift out of use.