Re: [Usability] Usability?

[Calum Benson <Calum Benson Sun COM>]

On 17 Feb 2006, at 20:36, Hynek Hanke wrote:

> Can someone please explain to me the metodology of this group?
> Is it people sitting and thinking and proposing what is best
> for the ``canonical user'' or is the group trying to get feedback
> from real users (request features, confusing things etc.) and
> fix these things?

It's a good question, and always worth discussing now and again :)

Of course, our aim should always be to get feedback from real users,  
first and foremost.  (And by that, I mean during the design stage if  
possible, not just afterwards.)  Inevitably though, because of the  
way open source communities work, that doesn't always happen or isn't  
always possible, so developers sometimes have to use their best  
judgement.

The GNOME Usability Project, and the GNOME HIG that it produces, are  
resources that can be called upon to help to guide that judgement.   
This list is another, although there aren't as many usability  
practitioners or researchers regularly contributing here as there  
used to be, so discussions can sometimes tail off without any  
informed resolution, which is unfortunate.

The usability study videos and reports at http://betterdesktop.org  
are another great resource, as are the other occasional GNOME  
usability studies run by Sun and others.

The best place to comment on usability issues with specific  
applications is bugzilla-- cc'ing usability-maint gnome bugs   Most  
maintainers will treat usability bugs as seriously as functionality  
bugs these days, which is a Good Thing.

> Also, can you explain the concept of the ``canonical user'' to me?
> Because, strangely, I don't mean them. (which doesn't of course mean
> they are not most Gnome users, I don't know...). Do you have some
> data about real users, which would show there really is something
> like the ``canonical user''?

There is no one canonical user for a desktop product, but we  
certainly have some reasonably well-defined target audiences that we  
need to bear in mind when designing any part of it.  This might be  
best done by writing up some personas[1], which would act as another  
one of those resources for helping developers to make good judgement  
calls... this is something we've talked about for a long time, but  
we've never quite managed to pull it together :/

> Also, what is the method for solving such usability problems?

Well, that's a whole subject in itself :)  Lots has been written  
about it, but frankly I'm not sure any open source project is a  
particularly good role model as yet.

In an ideal world, you'd run some usability studies and/or focus  
groups, identify the problem, involve representative users in some  
iterative prototyping to achieve the redesign, and run some  
comparative usability studies afterwards to make sure you'd improved  
things.

In GNOME, what mostly happens is that a bug is filed, there's a bit  
of discussion (hopefully involving the original reporter, and  
somebody from the usability project, but that's not always the case),  
and a consensus is reached about how to fix it.  Obviously there are  
some potential pitfalls here:

- most users outside the active community never file a bug (less than  
10% on most big projects), so we might be focusing on entirely the  
wrong things
- when a user from outside the active community /does/ files a bug,  
chances are we know little or nothing about them, how representative  
they are of the users we're trying to accommodate, or how many other  
people are having the same problem and haven't reported it (or don't  
know/care enough to do so)
- redesigning by consensus sometimes leads to compromises that don't  
necessarily improve matters

> Because I imagine they mostly lie in different Gnome components
> or even result from a combination of different parts of Gnome.
> Is there someone working directly on usability or are such
> issues reported to the maintainers of the software?

Issues that are specific to one application are best reported in  
bugzilla, as I mentioned above.  Issues that affect multiple  
components are usually better discussed on this list, initially at  
least, or might be more suitable for desktop-devel-list if they'd  
have a more fundamental/architectural impact.

Cheeri,
Calum.

[1] http://www.cooper.com/newsletters/2001_07/ 
perfecting_your_personas.htm

--
CALUM BENSON, Usability Engineer       Sun Microsystems Ireland
mailto:calum benson sun com            Java Desktop System Team
http://blogs.sun.com/calum             +353 1 819 9771

Any opinions are personal and not necessarily those of Sun Microsystems