Useful errors in technical translation?

Apr 14, 2016

Nobody wants incorrect translation. Too many errors in delivered texts can only compromise the translation supplier and destroy trust in a customer-LSP relationship.

There are, however, certain types of errors that, if they occur, can promote the translation quality, clarify terminology use by pointing out mismatches and inconsistencies in translation resources (glossaries or translation memories).

Errors to watch

There are several critical areas that should be analyzed and reported by translation suppliers as a part of follow-up on everyday projects:

  • dismissed glossary proposals
  • ignored fuzzy matches
  • unused exact matches
  • glossary recommendations

Such post-analyses as well as comments and recommendations from the translators who handled particular projects help to keep translations and translation resources error-free and in line with clients' needs and terminology.

For example, if a translator didn't use a proposed glossary expression, he should report a reason why it was not used (such as if a glossary proposal does not fit context, spelling/grammar is incorrect, the proposal is in the wrong language or the translation incorrect, etc.). Similarly if the translator has ignored a fuzzy match (due to differences not relevant in particular target language, etc.) this should be reported as well.

Such issues then become valuable feedback to content managers.

Get follow-up reports!

Comprehensive follow-up reports should be generated for each translation project handled by your translation supplier. These reports are supposed to include the information mentioned above, including glossary and fuzzy match issues. As a result, using these reports it will be easier to keep glossaries, translation memories and other linguistic resources up to date. And compared to case-by-case handling of translation issues, this is a systematic approach that perfectly fits any quality management system.