Hi,
I discovered with interest that your Google Translation Hub advanced tier offers document post-editing features, and, as part of that, includes an MTQP quality prediction score on a segment by segment basis.
This would be a very interesting feature to include in Cloud Translation API, particularly for TMS and CAT tools like Trados/MemoQ/Memsource that could then provide that information to the translator, similar to what a fuzzy match is for traditional Translation Memory technology.
It could also be very useful to decide whether a raw machine translation process (without review) is suitable for a document, or to identify the few segments that absolutely must go through human editing depending on your acceptable quality profile.
So my question is whether Google is looking at making this available in the API, or whether Google is treating that as proprietary information that you guys do not want to make available outside of your Google Translation Hub? I really hope the answer is the former, not the latter...
Thank you.
Michel Farhi
Principal Localization Engineer
NI (formerly National Instruments)
LinkedIn
Twitter
