As long as the target tags did not exist, I’m happy. I would not expect a bot to do semantic checks.
I would expect a bot to do reasonable structural checks, as already done by the name copying bot. For example, avoiding copying a name that has no Hebrew letters to name:he and similarly for Arabic. It could be improved, however, to also avoid copying non-textual “name” tags to “name:en”. Perhaps another bot can fix the 1000+ existing such duplicates.
@zstadler: I’ve thought about it and I agree with your view regarding autofixes. Detecting new user errors is more valuable than copying a few dozen extra names.
I still think I need to print logs or add fixmes, so that a human is more likely to fix the mismatches.