Re: [External] : Re: Concern about new binding.

[Thibaut Verron]

2021-02-05 9:16 UTC+01:00, Eli Zaretskii <eliz@gnu.org>:
>> From: Richard Stallman <rms@gnu.org>
>> Date: Fri, 05 Feb 2021 00:49:22 -0500
>> Cc: eliz@gnu.org, kevin.legouguec@gmail.com, stefankangas@gmail.com,
>>  spacibba@aol.com, emacs-devel@gnu.org
>>
>> More than that.  Over that time, how often have people
>> asked for such a global binding?
>
> We never bother ourselves with such questions; never did.  We consider
> ourselves to be aware and familiar enough with the Emacs usage
> landscape to make such decisions without polling users on each and
> every step, because doing so would slow down development to an
> unbearable crawl.  I always believed that at least part of the reasons
> we were nominated as maintainers was that people trust us to be
> capable of representing the bulk of Emacs users, and do it well enough
> to avoid too many serious mistakes.

I think this hits the nail right on. Some of us (myself included)
believe that this change underestimates how many Emacs users do not
consider C-x g to be a free-to-take binding.

> In a case such as this one, when one of the maintainers says "this
> makes sense", I expect to hear technical arguments for or against that
> (btw, only agreements were heard when the original decision in this
> case was made), but I do NOT expect to hear "go ask the world because
> you don't really know what you are talking about".

Without going as far as making a formal poll, I don't think it's
unreasonable to be as careful for binding a new key as we are for
rebinding an existing key.

This community has achieved a bit of a "conservative" reputation on
the latter, which may explain the surprise at how apparently
light-handed the same decision can be taken for a "free" key.

Besides, technical arguments were also brought forward: making
revert-buffer too easy-to-reach is dangerous, modes which need
frequent revert-buffer already bind it (directly or by inheriting from
special), there is auto-revert-mode, binding free keys will
necessarily break some users' configuration, the chosen key conflicts
with a major 3rd party package in a way which will break its users'
configuration.

As far as I can tell, the suggestion of a poll was only metaphorical,
as in "do we really need this in view of the drawbacks?".

> In all the 30 years of my uninterrupted active involvement with Emacs
> development, I don't remember even a single instance of polling users
> before making user-visible decisions.  (I may have missed one or two,
> but it cannot be more than that.)  I'm astonished to hear such demands
> now.  If this is indeed what's required from Emacs maintainers, I will
> seriously consider resigning, because I cannot in good faith support
> such ridiculous development practices, let alone such level of
> mistrust towards my and Lars's experience and knowhow.

I can only speak for myself, but I absolutely trust that all
maintainers know the Emacs community far better than I do.