KFeedbackWizard - another try

[Martin Ellis]

On Tuesday 13 Sep 2005 15:48, Charles de Miramon wrote:
> I don't think this kind of feedback wizard is very useful :
> 1) There is a lot of sociological theory that proves that multiple choices
> questionaries gives back a distorted version of truth. If we translate the
> questions in multiple language the problems become even worse due to
> cultural differences.

I'm sure something approximating good feedback is better than no feedback.

> 2) There are already plenty of channels (mailing lists, bugzilla, irc) of
> interaction between the developpers and our users.

People are always talking about entry barriers.

Using a communication medium that requires a good degree of technical skill 
ensures we only get feedback from technical users.

Bugzilla requires a log-in and is actually quite difficult for people who 
aren't familiar with how that kind of thing works (I've even seen technical 
people struggle with this.  Heck, I struggle with it sometimes).

Requiring that users remember a password for us seems a bit much. 
Remember we want *their* feedback.  This is about trying to get users to do 
*us* a favour.  Let's not make that difficult.

Mailing lists can also be quite difficult to manage for some people.  How to 
deal with that volume of mail?  Which list?  How do I subscribe? Am I allowed 
to post?   But we should really be making it *very* easy to provide feedback.

> 3) Who is going to sort and interpret the results and digest them for the
> developpers ?

> What would be much more useful is a voluntary feedback system of how people
> use kde :
It *is* voluntary.  It's not like they have to use the feedback wizard before 
they're allowed to use the program. ;o)

> - What configuration options they have used or not
> - What applets they have in kicker
> - What applications they have launched last month
> - What documentation they have read...
The questions are configurable per application.  Which is good, because I 
reckon that most application developers aren't even faintly interested in 
what kicker applets their users have.

> Having some kind of statistical engine telling us what is popular and
> unpopular would help next time we need to clear some dead wood in KDE, or
> make it possible to find some patterns of use and adapt our defaults to
> these patterns, or understand what is easy to find in KDE and what is badly
> explained or publicized

Perhaps.  But if people are annoyed by a feature so much that they write even 
one sentence about it in a feedback wizard, then that's the kind of thing 
that can be addressed directly, without having to think about numbers or 
statistics, or dubious interpretations thereof.

Martin