Ivor Hewitt ivor at ivor.org
Fri Nov 11 09:52:51 GMT 2005

Richard Moore wrote:
> On 11/10/05, Michael Pyne <michael.pyne at kdemail.net> wrote:
>> And to add to all that, consider this statement:
>> "Fully *half* of all people have *below* average intelligence!"
> And most people have more than the average number of legs (honestly,
> they do - think about it).
> More seriously: I wrote a tool that lets you measure how much the
> different actions get used, see http://www.kdedevelopers.org/node/376
> why not actually measure this stuff rather than guess?

Whilst such tools are interesting (and fun), they are of more use to 
monitor usage patterns and spot where users are performing actions in a 
sub-optimal manner and then try and find ways to make the best solution 
the obvious one (simple example such as running a M-x command in emacs, 
if it has a keybinding it will let you know in the status bar).

Simply saying hey this tool isn't used much so there's no point having 
it really prominent is too simplistic, but unfortunately a common 
reaction by a lot of software developers. To take another 
(over)simplistic example lets say we have a "recover deleted item" 
option for something, we analyse usage just by frequency and say hey 
this is never used even once my most users let's pull it out into a 
separate recovery application. The one time someone does need to use 
that feature they'll be very unhappy.

The only real way to get usability right is to run usability labs, sit
with a whole cross section of users running a series of tasks and
analyse the results.... even then you have a "schroedingers user"
problem :) where people try and find problems because they think that's
what's expected of them and they don't want to feel that they're wasting
peoples time.
This is much easier for a company with management control and a strict 
product release cycle to achieve than for open source projects.

The next best option is for everyone to eat their own dogfood (it's also 
important to try the catfood too - personally I've been running Gnome 
for the past two weeks to see if there is anything interesting - and 
Vista too, some interesting changes, although its hard to assess since 
it's as buggy as hell) and actively point out and help where usability 
is lacking.

Similarly however, suggestions and criticisms when raised by users need 
to be welcomed and considered, unfortunately this is an area I think KDE 
currently is not great at.

Usability is hard.


More information about the kde-core-devel mailing list