Re: Representation of the Emacs userbase on emacs-devel

[Stefan Kangas <stefan@marxist.se>]

Dmitry Gutov <dgutov@yandex.ru> writes:

> When we look at the polls about indentation style preference (where
> "tabs" can be as high as ~30% for certain languages), they don't prefer
> the kind of tab-based indentation that Emacs does. Which really means we
> only satisfy some tiny fraction of the users OOTB in any language.

I do appreciate the work Dmitry put into Bug#20322, which as far as I
can tell shows that spaces-over-tabs is much more widely used than
tabs-over-spaces.  The reason we didn't flip the default was that John
Wiegely was worried that it would lead to problems for someone out
there, and that the benefit of the change is too low.  I can see much
sense in that.

At the same time, since the data we have suggests that
spaces-over-tabs is more common/popular, indent-tabs-mode to t is an
unnecessary frustration for most users that we might want to remove.
One way we could move forward with this, or other changes, is to make
an experimental change on master.  That is, we would tentatively flip
the default on master after the emacs-28 branch is cut, and wait to
see what kind of feedback it results in.  If we don't get any reports
of breakage before Emacs 29.1, even after our pretests, the new
default wins.  Or some variation on that formula (we take the decision
after having had it for N days, etc.).

I think we have made such experiments only once (or twice?) so far.
I'm not saying we need to do this for indent-tabs-mode specifically,
but the process is there and could be used in more cases.  It could
even be used for changes we feel more confident about (i.e. the ones
that we today would just make, and then it stays because the decision
was already taken).