Re: Changes for emacs 28

[Ricardo Wurmus]

Göktuğ Kayaalp <self@gkayaalp.com> writes:

> On 2020-09-10 14:20 +03, Ricardo Wurmus <rekado@elephly.net> wrote:
>> At that point their initial enthusiasm has all but disappeared and they
>> glance over to their colleagues who use Rstudio or Atom (in 2017) or
>> that proprietary editor Sublime.  Everything seems so easy and
>> approachable and just as extensible.  They see their colleague use Git
>> from within Rstudio and wonder if they’d ever get to that point if they
>> will first have to configure Emacs to do all the basic things first.
>
> IMO we should draw a line between attracting users into a detrimental
> experience for them and making Emacs more approachable.  Rstudio is a
> _great_ fundamental project, and if the rest of my life wasn’t inside
> Emacs, I’d be using it these days as I’m studying R.  Org mode is better
> than Rmarkdown in certain respects, but if it’s just ‘do statistics in a
> nice environment’, I’d definitely recommend Rstudio over Emacs (or
> anything else, for that matter).  There’s no way Emacs will ever be as
> good at that particular task set, especially if that’s the only task set
> one will need it for.

In my work environment R is essential, but it’s not nearly the only
thing people have to use regularly.  A lot of work is done in Python and
with Jupyter, and RStudio is almost useless for those tasks.  There’s
also a lot of command line drudgery that could be done more conveniently
with “M-x shell” (editing the output of a tool directly), etc.

So there is a palpable desire for better tools to avoid the constant
context switching, but for every task people end up on the hill of a
local maximum, unable to reach something better (presumably Emacs),
because it doesn’t seem worth the effort to climb that mountain if you
first need to descend into the valley of inscrutable configuration.

-- 
Ricardo