Balasubramaniam L. wrote:
I face the same issue with the QA process in Across. In Hindi, the verb-subject positioning often becomes opposite to that of English, which means if a number occurs at the start of the sentence in English it will move to the end of the sentence in Hindi. The QA process does not recognize this, and invariably points this out as a number-mismatch error, even though the highlighted numbers in the QA window are exactly the same, but are in different parts of the sentence. This also happens when I type the numbers in the Hindi font (Mangal, for example). The numbers looks exactly the same on screen as the English numbers, but the QA thinks they are Hindi numbers and are different from the English ones, and flags them.
As Lincoln suggests, CAT tools are still dumb, and it is best to ignore such errors.
[Edited at 2016-08-13 04:05 GMT]
But your suggestion is no solution to the problem because if I ignore this type of warning, I might ignore actually missing/extra numbers.
[Edited at 2016-08-13 06:53 GMT]