Re: Proposal for an Emacs User Survey

[Thibaut Verron]

> Spacemacs is configuration, it is also considered software, so also
> for this is obvious that 20,000+ Microsoft Github users have given
> star/like on Github, as that is what I understand when I read 20.7k+
> -- so it is obvious, that need not be asked, asking would be useless
> compared to obvious statistical results. (Note, when I use the word
> Microsoft in front of Github, it is used in the sense of warning.)

I would think that those 20k are a gross underestimate, more on that
below. (And I understood that.)

> Unless Github statistics are fake, GNU Emacs developers can already
> think why those 20,700+ users are using Spacemacs and think how to
> improve Emacs in general. In my opinion there is no need to ask what
> is obvious. There is need to act.

The actual number is not obvious, beyond "at least 20k".


> I know what you mean, especially for the reason that I have been doing
> surveys with people on the phone, in person face to face, on the
> street, and by visiting people at their homes, I have got paid for
> making surveys, and I know how to evaluate surveys, and how to write
> them properly.
>
> My answer to that proposal is same as Drew's, something like
> implementation of the above menus in the Help.

So only free form, no multiple choice, and permanently in Emacs?
That's fine for me.

The goal is then sensibly different from that of the suggested survey,
and as you say, they are not mutually exclusive.

> My opinion is maybe in that sense more "pushy" then what RMS wants, I
> think that free software philosophy has to be pushy 10 times more than
> it is now, as the increase of proprietary software and troublesame
> abuses of people on this planet increased as well.
>
> > Okay, that is a valid point (although I would be interested to
> > know if people still use Marmalade). But then we might as well drop
> > the repository question altogether.
>
> I would drop all the questions, and just have it embedded in Emacs
> Help menu for people to complain, suggest improvements, etc. System
> already exists, but "Report Emacs bug" does not incite suggesting
> improvements.

Questions are useful to avoid writer paralysis though.
There could be three systems: "answer survey" (with a few free-form
questions to give inspiration), "send suggestion", "report bug".

>
> Of course. I would not trust the Microsoft Github data at all, and it
> does not show logically to be true, here is short analysis:
>
> - we assume here that Spacemacs has 20700 stars/likes on Microsoft
>   Github, because it says so on the page. Does that mean those are
>   "users" of Spacemacs? Hard to believe, see below why.
>
> - if I open the main Spacemacs Github page, I can see that last
>   improvement on Spacemacs was 29th February 2020, there is not much
>   going on. But Spacemacs is appearing as first on Github for the
>   search word "emacs".

Second is the emacs-mirror (and the first result for a google search
"github emacs" for me).

I don't know how github sorts its results, it is not just stars.

> - I can see number of contributors, number is 562 contributors, there
>   are so many contributors but the last update was 29th February 2020,
>   maybe I am mistaken. So many contributors are counted as such if
>   they report a bug, they do not write software necessarily. I can see
>   very unrelated or mixture of bugs reported:
>   https://github.com/syl20bnr/spacemacs/issues?q=is%3Aissue+is%3Aclosed

Some branches are more recent, see for example buffertabs, last
updated 2 weeks ago and 4k commits ahead of master.

Apparently Github defines contributor as "has published to master"
(possibly via a fork and pull-request?).

The list of contributors is not visible beyond the top 100, but those
100 all have at least one commit on master.

> It is untrue that 20700 stars mean there are so many users, people
> click how they want and wish. Spacemacs is definitely not for
> beginners, it is distributed from Git, so user must have and know how
> to git pull or clone it.

Are you sure about that? The website spacemacs.org has a big
"download" button which serves a zipped copy of master.
Does running spacemacs require knowledge from git after that point?

Regardless, it is possible to run spacemacs without ever visiting the
github page of the project, and github stars cannot track those.
So even though some of those 20k stars are probably not users, or are
no longer users, I still believe that the actual number of users may
be way more than 20k.

A properly conducted survey seems like a good way to get a better
estimate of the actual number.
But I agree that it is not a trivial task.