Published July 14, 2021 | Version 1.0
Video/Audio Open

SSHOC Demo accompanying Deliverable 4.9: Guidelines on the Use of Translation Memories in Survey Translation

Creators

  • 1. GESIS

Description

In a short demo “Use of Translation memories in survey translation” the CAT tool MateCat and the integration of public and private Translation Memories are introduced. The recorded video includes the following steps:

  • Adding the public TM to a project
  • Adding multiple private TMs to a project, including the Multilingual Corpus of Survey Questionnaires
  • Private TMs from other users / sharing private TMs
  • Importing TMX files, exporting, and deleting TMs
  • Using the TM key
  • Using TM matches during the translation process: pre-fills, locking, exact and fuzzy matches, context matches, auto-propagation

A few key terms from the demo shall be repeated here:

Matches in the TM can provide exact, in-context, or fuzzy information about previously translated text segments. While an exact match means that the entire segment of the source text for translation matches a source segment from a Translation Memory, an in-context match means that, beyond an exact match, the segments above and below are an exact match, too. Fuzzy matches include segments that have a sufficient amount of common content between a segment stored in the TM and the actual source text so that they can be useful as a starting point for translation. The threshold for usefulness must be defined by a translator during the translation task. Theoretically, a fuzzy match can have any value between a 99% and a 1% match. In practice, the default value for presenting fuzzy matches to the user in many CAT tools is between 65-70%; below this threshold, matches will not be shown unless a user changes the CAT tool settings.

During a translation project with several translators, a private TM can be used collaboratively if it is shared with others through a special private TM key. The special TM key is identified by a unique alphanumeric value. The key works as a password that enables collaborators to access the private TM. By using the private TM, the storage of translated and aligned text is visible exclusively for the creator of the TM and the person the key is shared with in contrast to the public TM, where every user can save their translations and overwrite the existing ones. For more information on how to share a private TM key with others, please refer to MateCat Post-editing and outsourcing made easy.

Files

Use_of_translation_memories_in_survey_translation_deliv4.9.mp4

Files (257.0 MB)