Re: "Why is emacs so square?"

[Po Lu]

조성빈 <address@hidden> writes:

> Skimming the Emacs Lisp intro doesn’t make one reasonably proficient in
> programming Emacs packages though. For one to understand how Emacs works
> (with it’s obscure naming scheme like windows), one has to jump a lot of
> hoops, and I can guarantee that one who is familiar with Algol-family
> languages can pick up the VSCode API much faster than picking up Lisp,
> Emacs, and the Emacs API.

This is what gets interesting: Emacs Lisp is a language in it's own
right, and the "Emacs API" is the Emacs Lisp language.  Emacs Lisp is
also a rather small and simple language.

You don't have to pick up "Lisp", or the Emacs "API", you only
pick up Emacs Lisp.

> But the default Emacs doesn’t have that, and that’s the problem.
> Which means that, for one to have nice OOTB experiences, one has to have
> a really good reason to use Emacs (like learning Common Lisp), then google
> how to configure Emacs, then encounter Spacemacs without knowing anything
> about evil or helm or ivy. And proficient Emacs users usually recommend
> not using a starter kit in the internet. (That’s my experience on trying
> to use Emacs.)

We could put a link to them somewhere, but that's something for RMS to decide.

> I don’t think OP was saying that we should use Electron for Emacs, but more
> that due to using Electron, it gives the stability that Emacs doesn’t give.
> Maybe you’ve only used Emacs on Linux, but at least on macOS, Emacs glitches,
> locks, and crashes very frequently, and that’s a non-starter for a lot
> of people.

If it "glitches, crashes, locks", that's a bug, and instead of treating
it as a fact, report it.  Plus, I know a lot of people who use Emacs on
macOS, and I even had to use Emacs on Windows a long time back, and
Emacs has always been rather solid.  The starter packs are also supposed
to work well, and it might also be a problem with your own config.

OTOH, Lisp code shouldn't be able to make Emacs crash (unless you're
doing stuff like running invalid bytecode, or overflowing the GC stack),
and if it does, it's also a bug that should be fixed.