On Wed, Sep 28, 2005 at 06:23:15PM -0600, John Taber wrote:
> cedric GEMY wrote:
> >Since few words remarks have been made on IRC about GUI, and since it
> >seems to be something important as shown in the M. Salsbury article, i'm
> >proposing to do a beginning of Inkscape UI analysis, as i've already
> >done for Scribus (http://www.le-radar.com).
> >It will take few days to organize it and after that may be 4 or 5 weeks
> >to get a first document.
> >If people have ideas, remark or anything to say about this please, go
> It's nice of you to volunteer to help out. However, the problem with
> most of these individual UI critiques is that they are 1) totally
> subjective and 2)based on one person's opinion. Even if the person is
> "experienced" or "expert" in UI (whatever that means) and even if a
> reader thinks comments are good, it doesn't necessarily reflect what
> most users think and certainly doesn't put any sort of priority on it.
> Most likely a real expert, like a PhD in cognitive science, would base
> opinions on statistical research.
> To be effective, UI analysis has to be quantified. You simply can't fix
> what you can't measure!.
> UI can be measured in several ways (here's a few):
> 1) Survey of "Feelings" - ie. 90% surveyed users had trouble with
> setting preferences in preference dialog or ranked the dialog poor.
> 2) Time - ie it takes an average of 20 secs to draw a straight line
> (including the mishaps). 0-2 secs = A, 2-3 seconds = B .....
> 3) Accessibility - ie swatches cannot be identified by a color blind person
> While the grading breakpoints are still somewhat subjective, it at least
> puts some quantification on the issues and stops the "I say, he says"
> differences, and it helps prioritize the major issues.
> So, maybe the best use of your willingness and expertise might be to
> develop a good UI survey and tally the results.
Ooh, a survey approach is a *great* idea. Cedric, what do you think of
creating a survey?
Here are some items which from my experience working with other users
may benefit from a usability survey:
* Installing Inkscape
* Advanced node editing, such as splitting paths into pieces, etc.
* Grids. Snapping to grid. Setting grid size and units, etc.
* Advanced text operations. Converting text to paths, breaking a piece
of text up into individual letters as paths, etc.
* Boolean operations. Generally due to confusion about the terminology
* Setting the background color to something other than transparent.
One idea for the survey could be to present the user with a picture, and
ask them to draw it, and then write down how easy or hard it was to do
using Inkscape's UI.
We could also ask them to time themselves by writing down the time when
finish start reading the question, and then again when they've finished
Anyway, what do others think?