bubbles8531

The QC'er who mentioned the cut lines said later that he believes they're fully visible in the backlog. (It does seem right that backlogs have their own text wrapping?)

I can say there is a lot. I had (in the past) been impressed with DeepL in working with some other languages (movie subtitle tl) and was curious how much of gap there is with VNs. I extracted the Japanese from the game scripts, built a glossary for proper noun translations (main/side characters and gangs), wrote scripts to reformat, ID and verify all 103670 lines going in & out of DeepL Pro (almost half of which are character captions..thus the discrepancy with the Tracker total), filled in all the missing H-scenes, repacked, and ""tested"" the ""translated"" game for playability (in-scene rendered text appears in the backlog, choice buttons are translated, nothing crashes...) Even with a surface text check... DeepL's misgendering problems (he vs. she) and the occasional utter translation disaster (which you won't even realize is wrong/misleading) made it clear there's just no comparison between the current state of AI and a human translator. To your point: It's my impression that even the most rudimentary native or non-native speaker could do better than DeepL in many cases.

I think I read August had a prior experience, once, releasing something before it was fully QC'd, and they got flak for that. That can be rough- I know from releases in a software field. But am I the only person who is a little worried about possible WW3, in not-so-far future? If it becomes a question of releasing nothing or something- I believe the work and man-hours of everyone involved deserves to be considered, as it could end up all a waste, which would be tragic. This is just my opinion. Anyway, I will keep on checking the thread and spreadsheet, and hope for the best