Re: Proposal for an Emacs User Survey

[Jean Louis]

* Adrien Brochard <abrochard@gmx.com> [2020-10-12 18:17]:
> Thank you for your feedback!
> 
> I've read many emails on this thread now and I think we've reached a
> point where some action needs to be taken. I have purchased
> emacssurvey.org and I will proceed to:

Well that is what I said, you can do it independent of GNU project.

> - digest all the feedback I got into a final list of questions

I hope you will use "check all that apply" method, and not have "one
option" among the choice of lists.

> - announce the survey opening date on as many channels as I can
> - build emacssurvey.org as a no-JS site such that everyone can visit
> - it

That is right, not hard to do, I have done very complex forms, there
are many ways how to do complex forms, I was using CGI and Perl, but
simple HTML could be just enough.

Always add: "Other" field, which is enough wide and welcoming, to help
people write what they wish else.

When doing professional survey you could not change the parameters
during the survey period, so changing questions would not be good. I
suggest keeping versions, you may do more versions in future.

I hope you will align the purpose of the survey with Emacs
development, so that development may find out what is number one
feature needed and wanted in Emacs, to improve Emacs that way.

> - on the survey start date, emacssurvey.org will have two options for
> submitting responses
>   - visit the questionnaire as plain text and email it to a dedicated
> address hosted on emacssurvey.org (which will be possible from
> Emacs)

That is great.

>   - use an online form which will potentially require non-free JS

It does not require non-free JS or any Javascript. It requires plain
HTML, easy to do, it is even possible to make the Org file, and
process by Emacs Lisp that in turn makes the full survey online.

If your survey page is more complex, and the more complex you make it,
the less results you will get, but if it is multi-page form, then
there are nice Perl modules handling the job. Just see
https://cpan.org and search for forms.

But using any other programming language to generate the forms and
capture its information in the database other than Emacs Lisp, would
be disgrace for me, so just do it in Emacs Lisp, it will perform well.

> - the survey will remain open for 6 weeks, plus some extra time for late
> responders

For 6 weeks survey, you need to put up some money to capture the
users, and if you advertise, you should disclose on which
communication channels you advertised, and evaluation should involve
that advertising.

Your advertising would need to run for exact time of the survey, like
6 weeks, and then the survey results are specific only for that
specific communication channels you used.

Let us say you advertise on Reddit, you would get quite different
results then if you advertise on other network, attracting for example
Github users, or if you advertise so that you attract Debian users, or
scientists, students in universities and similar.

> - I will post on as many channels every week to remind people who
> haven't responded yet (but obviously not share any partial results)

If you do a survey for which you still do various advertising during
the survey time, that cannot be just survey, you would need to reach
users quickly to answer quickly their questions. You could advertise
on some networks using keywords like "emacs" and reach 1000 users
within 24 hours, rather then posting links online for 6 weeks, as in
that way you are changing the dynamic of the survey and also
influencing it consciously or not consciously.

How you have advertised it, and in what time, you would also need to
take in evaluation, along with your personal experience or skills.


Jean