User:Seb35/Terminology

Random notes for adding a terminology module inside the Translate extension.

Big picture

• Terminology: set of keywords defined in a small context (a document, a specific topic area) with a special meaning; in the context of the translation, the terms have some fixed translation(s)
• The terminology database has to be created, either manually, either automatically, either semi-automatically, or it can be imported from other terminology databases
• It is wanted the translator is proposed suggestions of terms, possibly with a definition and sorted by likelihood (e.g. (ultimately) correct male/female/plural/grammatical case, correct letter case, etc.)
• Manage the input inflections of the words (grammatical forms) and output inflections

Backend:

• Efficiently match terms in existing and new messages, either on-line or off-line (e.g. during importation/creation of a new document to be translated), it has to be done only once per document

PoC

• gerrit:132384
  • Introduce a namespace Terminology: where content of the pages are the given terms to be searched
  • If a term is found in the Terminology: namespace, the translation used is the content of the page Terminology:$term/$language
  • Asumptions: the terminology database is already in the Terminology: namespace, no interface to add it
  • Problems: interferences between the languages (terms are search accross the whole namespace), inflections not managed, probably not scalable when the number of terms increases