[Kst] Usability measurement tool ?

[Brisset, Nicolas]

Hi,

I'm using and helping improve the kst plotting tool and I realize I
often make comments about the number of clicks / mouse travel / time
required to accomplish a given task to try and convince people to change
things. The keyboard may of course also be involved... I have the
feeling that it should be possible to evaluate a more or less obective
"cost" for some actions based on:
- number of clicks
- number of dialogs the eye has to scan correlated with their complexity
(number of widgets ?)
- number of inputs made
- distance travelled by the mouse 
- keyboard/mouse focus changes (let go the mouse to use the keyboard is
"expensive")
- cost of keyboard accelerators used (is there a sort of measure for the
accessibility of keyboard sequences, including modifier keys... and
taking into account internationalization as some keys are more or less
accessible according the keyboard layout ?)
- "pauses" (no user action) during a sequence, probably indicating
trouble finding something
- ...

I was wondering whether the approach outlined above has already been
formalized somehow and whether there exist programs (or plans to develop
one) that can measure this ? I know usability exercises often involve
observing users doing the task, but this requires quite a complex setup
and the logistics effort makes it hard to extend that to many users. An
automated tool may help gather more data easily and decide when there
are debates. Also note that the advantage of the approach is that you
don't necessarily need to implement a feature to test it as you can find
most values (even roughly) using paper/.ui mockups.
I started out thinking that kodo might be extended for that purpose
(hence the CC:), but I realize it may be a bit too much :-) I also
believe that the previous quantitative measures need to have different
weights for different user profiles (typically those with accessibility
problems, but not only), which can certainly be handled easily.

Note that I'm new to usability. Even though I have interest and
apparently a good feeling for it I have never taken courses nor gone
further than reading a couple of Alertbox columns (more Web-oriented,
though). So please pardon me if my ideas are stupid and don't hesitate
to give feedback :-)

Nicolas