- XCompose is based on "visual overlap", it's trying to emulate a key on a typewriter that would type without advancing. So for example if you type COMPOSE / o it becomes ø. This can easily type ü, ú, ǔ etc but the standard default XCompose file was designed in the 80s before unicode so it lacks ǘ. I have a custom XCompose file I copy between all my computers which is what makes this a "get around to it" kind of thing. But this fixes the problem for no one else

- In general, computer and phone eastern language input systems have diverged into slightly different evolutionary paths.

- I strongly suspect Apple devices either do, or can be configured to, have a reasonable unified behavior if you use the longpress method. I don't own Apple devices.

- I am using a slightly unusual input method called "XCompose", which is built in to Linux and can be enabled in Windows with "WinCompose".

Problems:

- There are ONE HUNDRED THOUSAND hanzi characters. You only need to know about 3000 in practice but 3000 is a large number.

- I alluded to this above, but the most common way to enter written Chinese into a computer is… to type the spoken Chinese equivalent phonetically!

The German language doesn't make a use of dual diacritics, but at least, publishers with experience with umlauts would not automatically neglect them just because they used to be hassly to typeset.

FWIW, Vietnamese uses dual diacritics extensively. A publishing house that puts out Vietnamese textbooks would likely have no trouble at all doing the same for Mandarin.

Maybe I should fiddle with the settings. Or switch to Keyboard Designer.

That said, how often do Germans *combine* the ü and ú diacritics? Like it seems like Germans would only be using the '` signs when speaking French, in which case they don't need the ü.

The "norm" that diacritics don't matter is pretty much unique to the American publishing practices and their overseas influencees.

- Resources that are likely to teach me Chinese words silently omit the diaeresis on ü

- However, I am then expected to know the diaeresis is there, because the Android IME which is the most convenient way to enter 中文 in a computer will expect me to substitute v for u, because Android expects I know 汉语 or else why would I be using a 汉语 phonetic input method?!

- The problem is international cultural norms downstream from a technological problem so I cannot fix it except with a time machine

And resources for 中文 speakers are, obviously, written in 中文; the only way 中文 speakers interact with Pinyin is computer IMEs. On these systems, you just replace u with v. Type "lv" to mean lǜ.

By now we have unicode diacritics so ǜ displays easy, but everyone's used to the compromises…