I haven't tried Fluent yet, but—in theory—could you not have both the message identifier and the original string in your code?

Their React example kind of points towards this: https://github.com/projectfluent/fluent.js/wiki/React-Bindin...

They provide a <Localized> component, which has an identifier and wraps a piece of markup containing the original:

  <Localized id="hello">
      <h1>Hello, world!</h1>
  </Localized>

Seems like the best of both worlds.

Per the docs the wrapped markup is to improve readability for devs, it isn't ever used. That seems like the worst of both worlds to me.

You may also be interested in this comment thread from the article page, where the author explains his opinion on Zig's approach to allocators:

https://www.rfleury.com/p/untangling-lifetimes-the-arena-all...

I don't use zig, but it seems to me that functions can choose exactly which type of allocator they want to accept (just by changing the type of the allocator parameter), making his point invalid.

This is somewhat tautological, since consonants and vowels are already speech sounds.

No, not really. The word "user" starts with a consonant sound, for example, even though "u" is a written vowel.

From what I have seen, people writing about linguistics do not usually bother specifying this distinction in this context. It is usually implied that an expression like "following a vowel" or "following a consonant" has nothing to do with orthography, but the actual phonetics. The orthography is not really considered a source of truth.

They are sounds as well as groups of letters, but the OP didn't specify, hence my clarification.

Some examples:

* "He has an MBA" because "M" sounds like it starts with a vowel (here: "em").

* "a university" ("you-ni...")