Users of modern systems obviously fail with the context problem. The translators of many websites either didn't know or didn't care whether "comment" and "reply" were used as nouns or verbs, just to take a really common example.

> 1. Is easily “fixed” by a very powerful feature called “search and replace.”

True, but reconciling translation files based on different keys can get messy. Using keys which aren't directly tied to the original language solves this.

> 2. You usually (or it’s recommended to) translate entire phrases

The kind of situation I'm thinking of is words that are used alone, but which are disambiguated by context -- for example, "clear" could be an action meaning "remove everything", or a synonym for "transparent", or an odd term for "approve".

Sure. This is how we do it at work. We have a function, let’s call it fixme(). When you want to change a string, say, “example,” to “sample.” You call fixme(“example”, “sample”). The implementation sees if “sample” == “sample” for non-English. If so, it uses the old translation. A few weeks later, we remove the fixme and old string. There’s a job that pings you about it in Slack so you don’t forget, IIRC. You don’t even need a pre-commit review to fix them.

Words used alone or need context should use the i18n libraries context abilities. Like p_(“clear”, “verb”) and p_(“clear”, “noun”). It’s also helpful to add comments and your translation system should be able to pick them up:

// translators: %d is an amount of money

If you are using a modern translation system that doesn’t support these features, perhaps “downgrading” to an older system is a good idea.