Some more analysis of the 2020 Emacs User Survey

Some more analysis of the 2020 Emacs User Survey

[TEC]

Hello everyone,

I thought there was a lot of untapped potential in the Emacs user survey
results, so I decided to do my own analysis.

You may find my work here: https://tecosaur.com/public/EmacsSurveyAnalysis2020.html [1].

This analysis was done with the intent of helping the Emacs community
understand itself better, and which aspects of Emacs could benefit the
most from development effort.

I hope that this may be of some interest :)

Timothy.

[1]: You can get a few more formats by changing the file extension to
     one of: .pdf, .org, .org.html.

Re: Some more analysis of the 2020 Emacs User Survey

[Jean Louis]

* TEC <tecosaur@gmail.com> [2021-01-07 22:39]:

> What’s with these modeline people? Disable the modeline but keep
> everything else. They’re crazy.

That is funny and nicely personalized statement.

That is good that you made different analysis with intersections,
describing things like:

> TUI is massively more popular with Vanilla users
> GUI very slightly more popular with Doom/Spacemacs than the rest

That is what I meant with "intersections", to find what sets are
common to other sets.

> Doomers like Ivy, Spacers like Helm

Yes, nice expressions.

 > Error checking

 >    Most vanilla users don’t make mistakes 😉
 >    Everybody else is fairly similar (mostly flycheck, some rather confident individuals, and a small slice of flymake)

With vanilla I have no mistake, be it ice cream or cake.

Most of graphs until 1.3.1. I do not find user friendly, not easily
understandable. Your text statements are useful.

Framework graph shows that many new users are discovering Emacs
through Doom and Spacemacs, that I find useful information as maybe
those projects could be somehow supported by GNU, or promoted by GNU
project, why not, as long as they promote fully free software.

Maybe even, GNU could distribute Doom and Spacemacs versions directly,
or possible include them in GNU ELPA.

> 2.1 Org mode purpose and Emacs improvements and Emacs strengths

Great that you made presentation by the tag or word cloud.

https://tecosaur.com/public/figures/tm_emacs_strengths_word_cloud.svg

Maybe most important insight is the "Emacs learning difficulties"
https://tecosaur.com/public/figures/tm_emacs_learning_difficulties_word_cloud.svg

where it shows keybindings as main obstacle.

In regards to number of Emacs users, I have used Debian popularity
package and already found analysis, I believe number is few more
millions than what StackOverflow found.

You missed the point when you said "The respondents are predominantly
on Linux (65%), with most of the rest on MacOS (25%), then a slither
on Windows (10%) / BSD (2%). This is a huge Compared to the 2020
StackOverflow Survey, BSD is 20x more prevalent, Linux 2.5x, MacOS 1x,
and Windows 0.15x. " as Linux is kernel, while MacOS or Windows are
operating systems.

You have not even one time mentioned that Emacs is GNU project. I am
negatively surprised. The word "GNU" does not exist there even though
the piece of software you talk of is named "GNU Emacs" and not just
Emacs.

Would there be any chance that you change that page and provide
reference to GNU project, and original GNU Emacs website?

Jean

Re: Some more analysis of the 2020 Emacs User Survey

[tomas]

[-- Attachment #1: Type: text/plain, Size: 360 bytes --]

On Fri, Jan 08, 2021 at 03:32:33AM +0800, TEC wrote:
> Hello everyone,
> 
> I thought there was a lot of untapped potential in the Emacs user survey
> results, so I decided to do my own analysis.

[...]

Thanks! Nice analysis. I just skimmed it, but I'll give it
a more thorough read later.

You put a lot of work in it, thank you!

Cheers
-- t

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: Some more analysis of the 2020 Emacs User Survey

[Arthur Miller]

TEC <tecosaur@gmail.com> writes:

> Hello everyone,
>
> I thought there was a lot of untapped potential in the Emacs user survey
> results, so I decided to do my own analysis.
>
> You may find my work here: https://tecosaur.com/public/EmacsSurveyAnalysis2020.html [1].
>
> This analysis was done with the intent of helping the Emacs community
> understand itself better, and which aspects of Emacs could benefit the
> most from development effort.
>
> I hope that this may be of some interest :)
>
> Timothy.
>
> [1]: You can get a few more formats by changing the file extension to
>      one of: .pdf, .org, .org.html.
Nice work, thanks!

At least now we can rule out "windows like" shortcuts are important;
since CUA mode to be the least used.

Am I reading you correcnt when I interpret that TUI = Emacs in terminal?

I am surprised in that case, since I experience Emacs GUI window as a
much better terminal, but might be just me :-).

Re: Some more analysis of the 2020 Emacs User Survey

[Christopher Dimech]

There is no point in doing pairwise comparisons because the result is always symmetric.
You are really complicating simple things.  I disagree that the survey was very successful.

Whatever analysis one does, respondents were heavily biased and the documentation is only
good enough to those who already are quite knowledgeable.

---------------------
Christopher Dimech
General Administrator - Naiad Informatics - GNU Project (Geocomputation)
- Geophysical Simulation
- Geological Subsurface Mapping
- Disaster Preparedness and Mitigation
- Natural Resource Exploration and Production
- Free Software Advocacy


> Sent: Friday, January 08, 2021 at 6:55 AM
> From: "Arthur Miller" <arthur.miller@live.com>
> To: "TEC" <tecosaur@gmail.com>
> Cc: "Emacs developers" <emacs-devel@gnu.org>
> Subject: Re: Some more analysis of the 2020 Emacs User Survey
>
> TEC <tecosaur@gmail.com> writes:
>
> > Hello everyone,
> >
> > I thought there was a lot of untapped potential in the Emacs user survey
> > results, so I decided to do my own analysis.
> >
> > You may find my work here: https://tecosaur.com/public/EmacsSurveyAnalysis2020.html [1].
> >
> > This analysis was done with the intent of helping the Emacs community
> > understand itself better, and which aspects of Emacs could benefit the
> > most from development effort.
> >
> > I hope that this may be of some interest :)
> >
> > Timothy.
> >
> > [1]: You can get a few more formats by changing the file extension to
> >      one of: .pdf, .org, .org.html.
> Nice work, thanks!
>
> At least now we can rule out "windows like" shortcuts are important;
> since CUA mode to be the least used.
>
> Am I reading you correcnt when I interpret that TUI = Emacs in terminal?
>
> I am surprised in that case, since I experience Emacs GUI window as a
> much better terminal, but might be just me :-).
>
>

Re: Some more analysis of the 2020 Emacs User Survey

[Stefan Kangas]

Arthur Miller <arthur.miller@live.com> writes:

> At least now we can rule out "windows like" shortcuts are important;
> since CUA mode to be the least used.

I wouldn't be so sure.  Remember that we have a survivor bias in this
survey: CUA mode would be more important to those users that did not
stick around long enough to respond to the survey.

Re: Some more analysis of the 2020 Emacs User Survey

[TEC]

Arthur Miller <arthur.miller@live.com> writes:

> Nice work, thanks!

Glad you found it interesting :)

> At least now we can rule out "windows like" shortcuts are important;
> since CUA mode to be the least used.

I can't say I can come to that conclusion myself with the data currently
available. This is what I take away from the results:
- CUA mode is used by a small fraction of the respondents
  - both a bias towards emacs-involved developers and survivor bias
    would be at play here
- Keybindings were the most common term in the "Emacs learning
  difficulties" section
Crucially:
- We have no idea how many people were aware of CUA mode and how to
  enable it when they /just started/ Emacs.

> Am I reading you correcnt when I interpret that TUI = Emacs in terminal?

Indeed.

> I am surprised in that case, since I experience Emacs GUI window as a
> much better terminal, but might be just me :-).

I have heard (and found) that TUI emacs can perform better than TRAMP
(at least at the moment) over an SSH connection. Perhaps this has
something to do with it?
With the intake of Vim users, there could also just be a number of
now-Emacsers who are familiar with working in a terminal.

All the best,

Timothy.

Re: Some more analysis of the 2020 Emacs User Survey

[Tim Cross]

[-- Attachment #1: Type: text/plain, Size: 1889 bytes --]

On Sat, 9 Jan 2021 at 04:21, TEC <tecosaur@gmail.com> wrote:

>
>
> > I am surprised in that case, since I experience Emacs GUI window as a
> > much better terminal, but might be just me :-).
>
> I have heard (and found) that TUI emacs can perform better than TRAMP
> (at least at the moment) over an SSH connection. Perhaps this has
> something to do with it?
> With the intake of Vim users, there could also just be a number of
> now-Emacsers who are familiar with working in a terminal.
>
>
> The TUI interface can also be more consisent across platforms. Also,
depending on the setup, running emacs from a terminal can result in a TUI
version rather than the GUI version and given on some systems (like macOS),
starting Emacs from the dock has some added complexities for environment
settings (because apps run from the dock don't run inside your login
shell). Running inside the terminal as  a TUI instance tends to isolate
platform differences a little - for example, depending on which version of
Emacs you are running on macOS (macports, brew from source, etc), the
fullscreen/maximise functionality tends to vary when running as a GUI.
Running in the terminal as a TUI makes things a little more consistent and
if you do also run remote, everything is consistent.

Plus I've seen quite a few people who also like to use tmux. For example,
this is one setup I've seen for pair programming.

All of this means a TUI instance can make everything feel more consistent
when working locally and remotely and when working on different platforms.

Your point on tramp is also valid. I love tramp and use it a lot (I also
use GUI rather than TUI), but in some instances, tramp can be less reliable
or slower than running Emacs remotely. The downside with running remotely
is maintaining your emacs config, but having it in git can simplify that
somewhat.


-- 
regards,

Tim

--
Tim Cross

[-- Attachment #2: Type: text/html, Size: 2512 bytes --]

Re: Some more analysis of the 2020 Emacs User Survey

[Lars Ingebrigtsen]

TEC <tecosaur@gmail.com> writes:

> I hope that this may be of some interest :)

Yes; very interesting.  Thanks for doing this.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no