Re: "Why is emacs so square?"

[Po Lu <luangruo@yahoo.com>]

Stefan Monnier <monnier@iro.umontreal.ca> writes:


> Because of how it evolved, there is clear separation.
> And it's hard to retro-fit a distinction after-the-fact.
> But I personally do think of the Elisp world as split into various "layers":
>
> A- The core language itself.
> B- The core standard library.

Are you referring to `subr.el' and friends, or built-in primitives, or
subrs in *fns.c, or both, or all three?

> Great example: putting it in (B) makes sense, yet at the same time I'd
> put buffer-local variables in (A), thus breaking the layering.

Precisely.

> There are other implementations of Elisp, some of which are not tied to
> an editor, so it can make sense (see the paper Mike Sperber and I wrote
> for HOPL-2020 for an example).  

Of the many implementations that I see listed there, most are from the
"let's rewrite Emacs (in our favorite language / better)" camp, and do
include some (or most of) the editing primitives.  Of the others, I
don't really see Emacs Lisp, but a glorified MacLISP with optional
lexical binding and some Emacs Lisp-ish features.

I suppose we're operating on different definitions of what constitutes
"Emacs Lisp".