Re: Emacs design and architecture

[Eli Zaretskii]

> From: Emanuel Berg <incal@dataswamp.org>
> Date: Sat, 16 Sep 2023 01:58:42 +0200
> 
> When we speak of parallelism, i.e. the use of multicore CPUs,
> that is, true concurrency, are we talking Lisp parallelism for
> constructs that are parallel in nature, e.g.
> 
> (let ((a execute-on-core-one)
>       (b execute-on-core-two) )
>   (+ a b) )
> 
> or are we talking parallelism in the sense to make Emacs as
> a program make use of multicores?

We are talking about both.  Except that true parallelism doesn't care
which core will do what job, but leaves that to the OS, asking only
that each job be done independently of the other. i.e. usually by
different cores.

> The former, Lisp parallelism, would be amazingly cool, but
> maybe the latter would bring more actual benefit since one
> could, if modules were skillfully separated, do lots of
> optimization for their specific purposes and needs.

If you know how to do one of them, you also know how to do the other,
because Emacs is a Lisp machine.