Re: Draining the Swamp: A Technical User's Experience

[Maciej Stachowiak <mjs noisehavoc org>]

On 08May2002 08:45PM (-0600), Richard Stallman wrote:
>     If you find this statement surprising, I recommend reading some books
>     about usability and human interface design.
> 
> I'm sorry, I did not realize you meant only to cite general principle
> that there is some amount of trade off between number of features and
> ease of using them.  I don't think anyone denies that general
> principle, but since it doesn't tell us how much cost comes with any
> given extension, it isn't conclusive.

You are right. The best way to determine the actual cost is to apply
user testing, but since that is costly and time-consuming, the
second-best way is to apply experience.
 
> I thought you were making a much more specific claim about this
> particular issue.  The idea that studies show this specific change
> would cause a major inconvenience is what I found surprising.

My intent was not to argue the merits of any specific change, but
merely to point out that configuration options have a cost, and that
this cost must be carefully considered for every option.

In general, I think GNOME does not yet provide a convenient interface
for many important preferences, but at the same time clutters the UI
with unimportant settings. GNOME 2 does a better job of making the
cut, but obviously it still has some distance to go.
 
> Basically, how easy it is to find a particular configuration option is
> not such a crucial point.  If one has to look a few panels and see
> which one really does a certain job, that is inconvenient, but not
> disastrous. 

Actually, for the average non-technical user, it *is* a crucial
point. If they cannot find the option they want quickly, they will
likely give up. They are also likely to become frustrated, detracting
from their ability to work with the system as a whole and perpetuating
the unfortunate fear of computers that is so pervasive in our culture.

I've watched video footage of user testing that showed users failing
to understand a UI that seemed very clear to me, as a programmer, and
it was a huge wake-up call. Most people simply do not have the
patience or desire to fiddle with things that programmers or other
highly technical users do.

Good software should make users feel at ease, not frustrated. If we
want free software to be accessible to all users, not just the
technical elite, we need to take their needs into consideration.

> A typical user doesn't make many configuration changes
> per day.  If it takes a minute to look thru a hierarchical list of
> configuration panels (like what GNOME configuration has now) and try
> two or three of them, that is a small cost.
> 
> Being able to do the configuration changes you need is well worth the
> cost of having to look for the one you need.

I can't agree with this as an absolute statement. It depends on many
factors, such as:

- What proportion of users need to make a particular configuration change.
- How much these users will be inconvienced if they cannot make the change.
- How much noise offering that option adds to the interface as a whole
(Does it require one extra checkbox in an existing simple dialog? Does
it require a whole new menu hierarchy of preference dialogs?)

It is impossible to make the call on any given potential configuration
change without considering all of these factors, considering these
same factors for other options, and considering how they fit together
into a coherent interface.

Usability is a realm of shades of gray and tradeoffs, not black and
white decisions, so there are very few absolute statements we can make
about it.

Regards,

Maciej