I'm not so sanguine. I see it as you can be mostly wrong, mostly correct, or very close to correct, for the software that you write and the customers it serves at a given point in time . Any of those choices are fine, so long as you know what you are getting into.

So for example if you are writing a CLI utility for the European market, you could probably get away with doing very little, it would be an ugly experience for, say, German users, but it would be quick. You could go to the next level, eg, supporting umlauts and ß, and that would have a big payoff. You could probably stop there. But if you were writing a utility to do OCR on ancient texts, you'd have to go full-bore and support everything..and you'd still be wrong on occasion.

It's just a question of economic tradeoffs being explicitly judged and thought about, and not assumed away.