Παρακαλώ! Να'σαι καλά!

Σ'ευχαριστώ! Για ορθογραφικό έλεγχο δες το jinx από τον Daniel Mendler: github.com/minad/jinx Μπορείς να το χρησιμοποιήσεις με πολλές γλώσσες ταυτόχρονα. Απλά πρέπει να έχεις τα σχετικά λεξικά εγκατεστημένα.

​@@protesilaos Thanks for the information about jinx! I found that Emacs supports 'enchant' as a spell-checker out of the box, which is the engine jinx uses. It seems to be worth a try. :)

@awllower2 You are welcome! Yes, jinx is worth trying.

Good point!

You are welcome!

It turns out I can't. OS overrides Emacs key binding if I use the same keys both for OS and Emacs.

You are welcome!

I'm using 29.1, the function is still activate-transient-input-method.

@@goofansu I think you mis-understood my comment. I meant the function was called "transient-input-method" and is now called "activate-transient-input-method".

@@awllower2 Ah, sorry, thank you for the knowledge!

No problem. :)

Thank you for checking this!

Good to know! I wonder what the main constraint is. For example, emacs-rime has a demo in their README which looks similar to what Emacs offers out-of-the-box. I do not know any CJK to try it out, but I can imagine how a predictive typing model would help in this scenario as I would not want to type a long document while having to select a number each time.

@@protesilaos Being able to type the pinyin of several consecutive characters in a sentence, and then selecting a completion, gives the completion framework more context to suggest appropriate completions

There's also a package called pyim, which is pretty easy to setup. Does anyone know the pros and cons of pyim vs emacs-rime

@@protesilaos I am a emacs-rime user, it is wonderful and better than the vanilla emacs input. Something easily appreciable is that rime puts the candidates in a posframe that's closer to your cursor, that's less eye travel between your input and the candidates. Under the hood, yes the predicative model is important, and it should learn from the language pattern of the user to predict better what's coming. E.g. if I type 道 often enough, then next time when I type `d`, I should see 道 immediately without having to type the full `dao`. When it comes to character combinations, for an English analogy, if you type `hello world` often enough, rime knows to offer the completion of it when you only type `hl`. But the above are quite standard for any pinyin input, the vanilla emacs should have it too. An extra features of rime is that it allows the seamless insertion of occasional western words into the Chinese typing flow without the hiccups of having to toggle input method. This is an amazing experience if you use western words a lot in your Chinese. Rime also allows to narrow down the candidates by intonation, very useful. And rime is highly customizable (by a yaml file, not the emacs .el file). For example, if I use right arrow → often enough, I can customize rime so that when I type `ra`, I get the unicode symbol → as my candidate, among other standard Chinese character candidates. I am just scratching the surface of rime here, the power users of rime can really boost their efficiency and precisely specify their peculiarities (such as their dialects). One last thing, emacs-rime is *not* rime, it is just an emacs front-end of the rime input engine, which is cross-platform and OS-independent. I can use other rime front-ends to type Chinese in emacs, without emacs-rime. The best thing about emacs-rime is that it knows when I type M-x, I won't input Chinese!