Re: Representation of the Emacs userbase on emacs-devel

[Dmitry Gutov <dgutov@yandex.ru>]

On 03.09.2021 09:28, Eli Zaretskii wrote:
> That's only true for changes of the default behavior, and key bindings
> are examples of such a change, at least the way they are proposed.
> There was talk about introducing a minor mode which would then be free
> to make controversial changes, including key bindings, but no one
> stepped forward to write such a mode.  Which I think is a pity, given
> how easy it should be to do that, and how many problems and
> frustrations it could potentially solve.

Again you try to change any discussion of a change into an "addition". 
Something that wouldn't change anything in the default behavior.

> Once again: significant changes in behavior should generally be
> introduced as opt-in features, then the friction will be much lower
> and in the long run we could perhaps introduce changes at a faster
> pace.  Thus, people who insist on making changes in the default
> behavior are shooting themselves (and Emacs, if their opinions are
> right) in the foot, IME.

A lot of the time I'm talking about features that are already available. 
There's nothing to "introduce". But simply by having a wrong default we 
make Emacs harder to adopt with little benefit.

>> Or take indent-tabs-mode, an old pet peeve of mine:
>> https://debbugs.gnu.org/cgi/bugreport.cgi?bug=20322 I can talk about
>> contemporary practice, whole-industry polls, threads with personal
>> opinions anywhere, threads with people expressing confusion about the
>> current default behavior... but a few people say a change will be
>> inconvenient -- and it moves nowhere.
> 
> indent-tabs-mode is an existing option, so your insistence on turning
> it on by default in the face of resistance is ... peculiar.  We did
> turn it on in some of our sources.

Turning it off by default.

What's so peculiar about changing the behavior that flies in the face of 
existing practice in all programming languages out there? And which 
causes confusion in new users?

The way it is implemented made sense decades ago, but these days even 
users who want tabs for indentation are surprised by Emacs behavior in 
this area (because most of those users want 1 tab to mean 1 indentation 
level).