gathering data on how people use Emacs

gathering data on how people use Emacs

[Mike Mattie]

[-- Attachment #1: Type: text/plain, Size: 483 bytes --]

Hello,

With the recent discussion on how people use Emacs I have been pondering how to acquire use data that would illuminate
the discussion with fact.

Emacs has a nifty facility for recording the input events: these keystroke dumps, if suitably scrubbed of personal
data, and keyed to an anonymous ID could be used to populate a database that could be mined for the kind of information
needed in the discussions on user interface design.

WDOT ?

Cheers,
Mike Mattie

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: gathering data on how people use Emacs

[paul r]

2008/3/25, Mike Mattie <codermattie@gmail.com>:
> Hello,
>
>  With the recent discussion on how people use Emacs I have been pondering how to acquire use data that would illuminate
>  the discussion with fact.
>
>  Emacs has a nifty facility for recording the input events: these keystroke dumps, if suitably scrubbed of personal
>  data, and keyed to an anonymous ID could be used to populate a database that could be mined for the kind of information
>  needed in the discussions on user interface design.
>
>  WDOT ?


How about making a simple web page, with 3 fields :
 1) How many years have you been using emacs
 2) Paste your .emacs here
 3) Write here any suggestion about default settings

It would be easy to see what people activate in their .emacs, and in
which proportion.
A email confirmation mechanism would be welcome to prevent people from
being tempted to send 1000 differents automatically generated .emacs
with evertime (transient-mark-mode 1) in them :)

Re: gathering data on how people use Emacs

[Bastien]

I'd rather poll users with simple questions like:

- Should `transient-mark-mode' be ON by default?
- Should `show-paren-mode' be ON by default?

Of course, `transient-mark-mode' might not sound familiar to newbies.
Then we need to find a better way to describe what it actually does.

FWIW we can re-use a simple safe form like this one:

  http://www.legito.net/report_bug.php

The output of this form is sent by email and the content of the email
looks like:

,----
| * [#A] Bug report :BUG:
|   <2008-24-24 Mon 17:00>
|   :PROPERTIES:
|   :ID: fu3ouer3dt4igvt7fa2u4c6qv1
|   :Package: org.el
|   :Emacs: Emacs 22
|   :END:
`----

So that you can easily gather the result of the poll in an Emacs
org-mode file.

-- 
Bastien

Re: gathering data on how people use Emacs

[Mike Mattie]

[-- Attachment #1: Type: text/plain, Size: 2767 bytes --]

On Tue, 25 Mar 2008 02:17:19 +0100
Bastien <bzg@altern.org> wrote:

> I'd rather poll users with simple questions like:
> 
> - Should `transient-mark-mode' be ON by default?
> - Should `show-paren-mode' be ON by default?
> 
> Of course, `transient-mark-mode' might not sound familiar to newbies.
> Then we need to find a better way to describe what it actually does.
> 
> FWIW we can re-use a simple safe form like this one:
> 
>   http://www.legito.net/report_bug.php
> 
> The output of this form is sent by email and the content of the email
> looks like:
> 
> ,----
> | * [#A] Bug report :BUG:
> |   <2008-24-24 Mon 17:00>
> |   :PROPERTIES:
> |   :ID: fu3ouer3dt4igvt7fa2u4c6qv1
> |   :Package: org.el
> |   :Emacs: Emacs 22
> |   :END:
> `----
> 
> So that you can easily gather the result of the poll in an Emacs
> org-mode file.
> 

A poll is more of a political measure than a rational one IMHO. If the experts in this
issue are still negotiating terminology and the consequences of design on code and users
I don't see how a polling of people ignorant to the particulars will shed useful light
on the subject.

That doesn't mean that I object to the idea, merely that the data should be considered political
rather than rational. The same perspective discounts common sense as more dangerous than useful.
I struggle to recall a single moment in human history when common sense of the time had any
congruence to reality. In fact common sense is usually the cause of financial boom/bust cycles and other
embarrassments of the species.

Again, a poll is at least method, so I will support such a thing as progress, and vote if that
is to be the method for decision making.

I did, on my walk, strike upon an idea for facilitating the search for the Right Thing. We are, as a group of 
programmers trained in science and rational analysis. We should apply our skills and tools effectively. I think 
a good start would be leveraging the EmacsWiki.

There we could setup pages that define terms with acceptable precision, build analysis, and pool knowledge
until our conclusions and the methods by which we reach them are the pride of the community.

E-mail can be effective in debate but often knowledge is diffused throughout the thread rather than concentrated
as a central document does.

WDOT ?

I am willing to put time into a constructive collaboration in pursuit of the Right Thing, though my time is like
everyone else's limited. I will poke around on Apple's HIG research, see if I can extract some definition of terms
and research in this area. I am not a Wiki guy, so I would expect a Wiki enthusiast to flash some skill in getting
a initial structure set properly.

Cheers,
Mike Mattie

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: gathering data on how people use Emacs

[Richard Stallman]

Polling the users is not a matter of simply counting votes.
The most important part of the answer is "why".
What is the scenario in which a certain feature is helpful
or inconvenient?  What aspect of it is helpful or
inconvenient?

Thus, trying to use better technology to count votes is missing
the point of the poll.

Re: gathering data on how people use Emacs

[Mike Mattie]

[-- Attachment #1: Type: text/plain, Size: 444 bytes --]

On Tue, 25 Mar 2008 14:31:10 -0400
Richard Stallman <rms@gnu.org> wrote:

> Polling the users is not a matter of simply counting votes.
> The most important part of the answer is "why".
> What is the scenario in which a certain feature is helpful
> or inconvenient?  What aspect of it is helpful or
> inconvenient?
> 
> Thus, trying to use better technology to count votes is missing
> the point of the poll.
> 
> 

well spoken.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: gathering data on how people use Emacs

[Bastien]

Richard Stallman <rms@gnu.org> writes:

> Polling the users is not a matter of simply counting votes.
> The most important part of the answer is "why".
> What is the scenario in which a certain feature is helpful
> or inconvenient?  What aspect of it is helpful or
> inconvenient?

Fully agreed.  

I suggested simple questions, not simple answers.

> Thus, trying to use better technology to count votes is missing
> the point of the poll.

Yes.  The whole point of using the form I proposed is precisely to get
detailed comments and scenarii, not raw yes/no.

If people suggest a list of questions and associated fields for such a
poll, I'd happily draft a new one.

-- 
Bastien

Re: gathering data on how people use Emacs

[Don Saklad]

I like emacs for the control it gives over the text
when editing.

Finding the more obscured commands is too difficult.









.

Re: gathering data on how people use Emacs

[Mike Mattie]

[-- Attachment #1: Type: text/plain, Size: 2083 bytes --]

On Wed, 26 Mar 2008 00:32:02 +0100
Bastien <bzg@altern.org> wrote:

> Richard Stallman <rms@gnu.org> writes:
> 
> > Polling the users is not a matter of simply counting votes.
> > The most important part of the answer is "why".
> > What is the scenario in which a certain feature is helpful
> > or inconvenient?  What aspect of it is helpful or
> > inconvenient?
> 
> Fully agreed.  
> 
> I suggested simple questions, not simple answers.
> 
> > Thus, trying to use better technology to count votes is missing
> > the point of the poll.
> 
> Yes.  The whole point of using the form I proposed is precisely to get
> detailed comments and scenarii, not raw yes/no.
> 
> If people suggest a list of questions and associated fields for such a
> poll, I'd happily draft a new one.
> 

Here are some basic questions I thought of to give some context.
It's hard to ask directly for the information you want without loading the questions, so I fell
back on some basic background information as a starting place.

What was the motive or appeal that sustained your climb of the Emacs learning curve ?

What quality keeps Emacs indispensable ?

Years of Emacs Experience ?

Programming Experience: 0-9

Primary applications, or what do you use Emacs for pre-dominantly ? apps,web dev, books ?

Native language ?

Age ?

use of mouse ? 0-4

Most useful Modes ?

Least useful Modes ?

Size of .emacs ?

How do you go about learning a new Emacs feature you need ?

subscribed to emacs-devel y/n

subscribed to help-gnu-emacs y/n

systems you use Emacs on ? (autoconf triplets)

Emacs version ?

External Elisp packages used: (significant to daily workflow at least)

Most difficult command to learn ?

Easiest command to learn ?

command with best DWIM ?

command with worst DWIM ?

Use marks for ? (what is a mark should be a valid answer) 

favorite movie/book ? (throw at least one fun question in just so people don't get bored filling out a survey.
Could do the "top 5" for the fun question).

Cheers,
Mike Mattie

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: gathering data on how people use Emacs

[Robert J. Chassell]

    Here are some basic questions I thought of ...

    use of mouse ? 0-4

That is a bad question since there are times with mouse control of the
cursor when a sighted person should use it and other times not -- for
example, it makes sense when moving to a different and distant part of
the buffer that is visible on the screen.  

(When moving forward a character or word, it makes sense to invoke the
commands `forward-char' and `forward-word' with `C-f' and `M-f'
respectively.  You waste more of your life, that is to say, are less
efficent, if you move your hands to use either arrow keys or the
menu.)

Permanently and situationally blind people should not use a mouse; for
example, few people want other drivers to take their eyes off the road
to look at a mouse cursor.

    Most useful Modes ?

Another bad question, since modes should be used when appropriate.
Thus, when using Mail mode you should not be programming Emacs Lisp.
(You can use the *scratch* buffer for that.)

There are other poor questions in that list.  I would not use it.

-- 
    Robert J. Chassell                          GnuPG Key ID: 004B4AC8
    bob@rattlesnake.com                         bob@gnu.org
    http://www.rattlesnake.com                  http://www.teak.cc

Re: gathering data on how people use Emacs

[Mike Mattie]

[-- Attachment #1: Type: text/plain, Size: 1910 bytes --]

On Wed, 26 Mar 2008 11:28:58 +0000 (UTC)
"Robert J. Chassell" <bob@rattlesnake.com> wrote:

>     Here are some basic questions I thought of ...
> 
>     use of mouse ? 0-4
> 
> That is a bad question since there are times with mouse control of the
> cursor when a sighted person should use it and other times not -- for
> example, it makes sense when moving to a different and distant part of
> the buffer that is visible on the screen.  
> 
> (When moving forward a character or word, it makes sense to invoke the
> commands `forward-char' and `forward-word' with `C-f' and `M-f'
> respectively.  You waste more of your life, that is to say, are less
> efficent, if you move your hands to use either arrow keys or the
> menu.)
> 
> Permanently and situationally blind people should not use a mouse; for
> example, few people want other drivers to take their eyes off the road
> to look at a mouse cursor.
> 
>     Most useful Modes ?
> 
> Another bad question, since modes should be used when appropriate.
> Thus, when using Mail mode you should not be programming Emacs Lisp.
> (You can use the *scratch* buffer for that.)
> 
> There are other poor questions in that list.  I would not use it.
> 

Maybe you understood the purpose of the questions, but maybe you didn't. The idea of the questions
is to figure out how people are using Emacs. Personally I don't use Emacs while driving, but if
someone wanted to add that as a rationale for mouse = 0 ....

Some people even use Emacs to respond to a survey question list with tutorials :)

I would like to know what pieces of 3rd party elisp are most commonly used, the degree to which it is
used in practice, etc. The experience stuff is obvious. The mouse thing is important for how people
use Emacs.

Modes indicate applications: what people use Emacs for.

could you suggest some better/fairer questions then ?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: gathering data on how people use Emacs

[paul r]

2008/3/26, Mike Mattie <codermattie@gmail.com>:
> Here are some basic questions I thought of to give some context.
>  It's hard to ask directly for the information you want without loading the questions, so I fell
>  back on some basic background information as a starting place.

I think it is a good start. I can't wait to see what the shape of
distribution of age is.

>
>  What was the motive or appeal that sustained your climb of the Emacs learning curve ?
>
>  What quality keeps Emacs indispensable ?
>
>  Years of Emacs Experience ?
>
>  Programming Experience: 0-9
>
>  Primary applications, or what do you use Emacs for pre-dominantly ? apps,web dev, books ?
>
>  Native language ?
>
>  Age ?
>
>  use of mouse ? 0-4
>
>  Most useful Modes ?
>
>  Least useful Modes ?
>
>  Size of .emacs ?
>
>  How do you go about learning a new Emacs feature you need ?
>
>  subscribed to emacs-devel y/n
>
>  subscribed to help-gnu-emacs y/n
>
>  systems you use Emacs on ? (autoconf triplets)
>
>  Emacs version ?
>
>  External Elisp packages used: (significant to daily workflow at least)
>
>  Most difficult command to learn ?
>
>  Easiest command to learn ?
>
>  command with best DWIM ?
>
>  command with worst DWIM ?
>
>  Use marks for ? (what is a mark should be a valid answer)
>
>  favorite movie/book ? (throw at least one fun question in just so people don't get bored filling out a survey.
>  Could do the "top 5" for the fun question).
>

Re: gathering data on how people use Emacs

[Richard Stallman]

    > Thus, trying to use better technology to count votes is missing
    > the point of the poll.

    Yes.  The whole point of using the form I proposed is precisely to get
    detailed comments and scenarii, not raw yes/no.

I am sorry I misunderstood this point.