Why can't we decide on error handling conventions?

I see loads of blogs, conference talks, tweets, etc talking about *the answer* for error handling in rust. So clearly there's no consensus on what's "correct".

I understand that in different circumstances, different amounts of context and granulity of error types is required, but id have though we can work out a single way to do that flexibly.

So my question is, why all the disagreement? What more do we need to figure out? Are other languages doing this better?