Re: Entering emojis

[Eli Zaretskii <eliz@gnu.org>]

> From: Stefan Kangas <stefankangas@gmail.com>
> Date: Thu, 28 Oct 2021 16:37:55 -0700
> Cc: larsi@gnus.org, emacs-devel@gnu.org
> 
> But in the specific case of emojis, I would have expected that the
> emojis just get deleted in one go, as that is what happens in other
> software where I use emojis.  (Mostly chat programs on my phone.)
> 
> Would it be possible and make sense to handle emojis differently from
> other grapheme clusters in that regard?  The desire not to retype them
> from scratch might be less of a concern now that we have `emoji-insert'.

First, we don't yet have emoji-insert, it's only on master (which I
think is a mistake, given the significant improvements in this are we
will have in Emacs 28; but I digress).

And second, the current behavior could be quite useful in the Emoji
case as well.  For example, consider the case where you typed the male
version and you want to change it to the female version instead.  Or
when you copy-paste the Emoji from some other text, then want to
change it in small ways.

The behavior you suggest could be a user option, by default off, and
not specific to Emoji.  Changing the behavior unconditionally, or even
making that the default of that option, makes no sense to me, since
the current behavior is very old, and I never heard any complaints
about it.  making it specific to Emoji is possible, but sub-optimal,
because it would require the user to remember what kind of composition
sequence he/she is typing; and many users would have a difficulty to
know whether a given composed sequence is considered Emoji or not,
given that more and more non-Emoji characters can have "Emoji
presentation form" requested by appending VS-16.

And I submit that, as in many other situations in Emacs development,
we are jumping too fast to conclusions.  Here's a new feature, fresh
out of the oven, not even 100% complete yet, and we bump into behavior
we never noticed before, and immediately want to change it because it
surprises us.  Why not hold our horses until we have _some_ experience
with the new feature, and then consider this and other aspects with
some perspective in mind?  What's the rush to make changes in
unrelated functionalities just because we were surprised by a TIL-like
phenomenon, with hardly a few keystrokes of experience under our
belts?