Re: Emacs Lisp's future

[David Kastrup]

Nic Ferrier <address@hidden> writes:

> Lars Magne Ingebrigtsen <address@hidden> writes:
>
>> I think there's a somewhat greater cultural overlap between Emacs Lisp
>> and Common Lisp people ("get things done") than between Emacs Lisp
>> people and Scheme people ("interesting academically").
>>
>> Anyway, I think the dangled sorta-promise that Emacs would eventually
>> shift to Guile might have stifled Emacs Lisp development.  Whenever
>> somebody has brought up the issue of evolving Emacs Lisp (to
>> multi-threadedness or whatever's fun), they're usually discouraged by
>> others piping in with "oh, Emacs is moving to Guile, anyway, so don't
>> bother".
>
> I don't think that's true.
>
> From my perspective, what's stopping more people getting involved is the
> community, which is sometimes quite negative and the tooling, which is
> baroque.

Well, let's take a comparison:

git shortlog -s --since "1 month ago" origin/master

in a current GUILE repository gives

     2  Andy Wingo

while in a current Emacs repository it gives

     4  Alan Mackenzie
     1  Alp Aker
     4  Christoph Scholtes
     1  Christopher Schmidt
     5  Daniel Colascione
     1  Detlev Zundel
    27  Dmitry Antipov
    50  Eli Zaretskii
     1  Fabián Ezequiel Gallina
    26  Glenn Morris
     1  Ivan Shmakov
     5  Jan D.
     1  Jay Belanger
     1  João Távora
     1  Kan-Ru Chen
     1  Karol Ostrovsky
     3  Katsumi Yamaoka
     8  Ken Brown
     1  Ken Olum
     1  Lars Ljung
     2  Lars Magne Ingebrigtsen
     4  Leo Liu
    15  Michael Albinus
     1  Michael Heerdegen
    34  Paul Eggert
     1  Rasmus Pank Roulund
     1  Reuben Thomas
     9  Sam Steingold
    16  Stefan Monnier
     2  Thierry Volpiatto
     1  YAMAMOTO Mitsuharu
     8  martin rudalics

That does not particularly make Emacs look like a project keeping people
from getting involved.  Of course, the project/repository structure of
both projects is different, but the basic idea that forward-looking
development happens in the master branch is loosely common to both.

If we take GNU LilyPond, incidentally based on GUILE 1.8, for reference,
we get something like

git shortlog -s --since "1 month ago" origin
    60  David Kastrup
     4  James Lowe
     6  Janek Warchoł
     3  Jean-Charles Malahieude
     4  Julien Rioux
     5  Keith OHara
    19  Phil Holmes
     3  Trevor Daniels
     1  Walter Garcia-Fontes

Now this is a project that has to suffer from a lead developer who is
considered to be sometimes (or more) quite negative, and the tooling,
namely GUILE 1.8 which is considered outdated and unmaintained for
something like 5 years or so, can also be called baroque.

The result is a quite more peaked distribution of contributions per
developer than with Emacs which has sort of a plateau at the top.

So I think that the news of Emacs' demise due to the named reasons is
quite exaggerated.  There is always room for improvement, of course, but
no particular need to panic.

-- 
David Kastrup