Frequently, it is necessary to translate a large volume of documents in a short amount of time or to quickly grasp the essence of something. It is not surprising that machine translation (MT) is frequently considered for these tasks. There’s no doubt that MT can translate faster than humans and at a fraction of the cost. But, with no human intervention, important pieces of information can be lost or, worse yet, completely misrepresented.
When using MT, it is important to decide which is the desired quality level of the output. Generally, there are two levels of post-editing, light or full.
Light post-editing entails making the fewest possible modifications to the raw MT output in order to make the translation comprehensible, factually accurate, and grammatically correct.
Full post-editing, on the other hand, must produce translations that are grammatically flawless, employ the correct terminology consistently, have the appropriate tone and style, and are devoid of stylistic inconsistencies and variations.
The following table summarizes the differences between these two kinds of post-editing:
|Light post-editing||Full post-editing|
|Minor grammar and spelling errors are acceptable||Correct spelling and grammar|
|Punctuation variations and errors are acceptable||Punctuation is accurate and uniform|
|Spelling variations are acceptable||Spelling is uniform|
|Terminology is comprehensible and practical||Terminology is precise and consistent|
|Style and tone are not offensive||Style and tone are appropriate for the content|
|Variations in style are acceptable||Style is consistent|
|Formatting is not important||Specific formatting requirements|