Great question. Next week we're rolling out the ability to attach context to translation requests, which means sending descriptions and pictures of the requested translation to the translator. Would that solve the issue, or is there something else we could do to address this problem?