Re: How to poll the users

[Dmitry Gutov]

On 26.04.2020 06:25, Richard Stallman wrote:
> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>    > Speaking of the default values, what do you think about using your
>    > scenario to ask the users about their preferred value of
>    > indent-tabs-mode? I can create a poll with four options: "strong tabs
>    > preference", "strong spaces preference", "mild tabs preference", "mild
>    > spaces preference". People can add extra explanations in comments.
>
> The idea of a poll is to understand the effects of the various
> choices, and find out the specific reasons for users to prefer this or
> that, and evaluate their significance -- NOT to count how many people
> prefer each choice.

A numbering poll is still somewhat useful, if only to get the general mood.

But in the description we can ask everybody to actually write a comment. 
*Or* upvote one of the existing comments, if it reflects their 
experience accurately.

Both these options together seem to make responses over email (which 
would otherwise be invisible to the public and thus miss out on the 
chance to spur some discussion) unnecessary. But we could also encourage 
people to write emails, if you so prefer.

> It is a mistake to count the answers, because a given change can
> affect one user very often, and affect another user only rarely, but
> they might both state a "strong preference".

I think the time when it affects the users the most is just after Emacs 
is installed for the first time, before they know about this variable, 
and how to customize its value to nil. So it doesn't affect people 
"often", but it does affect a lot of people.