Re: Why is Elisp slow?

[<tomas@tuxteam.de>]

[-- Attachment #1: Type: text/plain, Size: 1778 bytes --]

On Thu, May 02, 2019 at 02:49:22PM +1000, Van L wrote:

>  ;; "No matter how hard you try, you can't make a racehorse out of a pig.
>  ;; You can, however, make a faster pig."

[...]

> If a big advance from an Industry player, such as FB, were to release a
> JIT, that the GNU Emacs legal dept. does not object to, does that help?

I, for one, appreciate Emacs much more for other characteristics than
its speed.

Its extremely judicious developer community, which manages advances
at a steady pace (in a 30 year old project) while generally avoiding
disrupting surprises for its regular users is... amazing.

I've yet to run into a serious limitation posed by its "slow" elisp
implementation.

It's the same outside Emacs. For most of my day-to-day problems, I
use known-slow language implementations. For those rare exceptions,
there are other tools.

Of course Common Lispers (and to some extent Schemers, and other
functional programmers [1]) are spoilt by extremely clever implementations,
and this is what the comment's author was hinting at -- but it takes
a wizard to master those tools. It takes a grand wizard to hack them.

The real feat here is that Emacs offers an environment which is still
amenable to non-wizards. In practice, I think, the biggest optimization
potential still exists in making contributors better than in making
the core implementation faster. In that metric, I'd say, Emacs succeeds
very well [2].

Cheers

[1] This is an example of what I'm talking about:
    https://dl.acm.org/citation.cfm?id=3276489

[2] Here I can just give that reference back to you
    "Evolution of Emacs Lisp" by Stefan Monnier & Michael Sperber
    https://www.iro.umontreal.ca/~monnier/hopl-4-emacs-lisp.pdf

-- tomás

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]