Re: [Usability]Feedback on GNOME 2


Several of your examples in arguing for more configurability cite
simple bugs or misfeatures, that could be fixed outright (or already
have been). I don't know why people assume that an option solves any
problem. "I want an option to have no segfaults!" ;-) "Unbreak this
application please!"

"It's buggy, I need more config options" (just fix the bugs?)

"The default sucks, let me change it" (just fix the default?)

"I don't know what people use or want or why, so let's make it an
option" (how do you even know what _options_ they want then?)

Magic prefs dialog pixie dust.

> Users want simple; but they also want flexible.  There is a reason
> that UIs, and computer programs in general, have developed
> increasing customization over the years; it's that it's what people
> want.  It's not possible to define a single UI behavior that will
> work for everyone.  People want choice.  The more, the better.

I don't think it's productive to have an argument about "prefs bad" or
"prefs good." As I argued on this page:, the interesting question
is where you are going to draw the line between prefs that you add,
and those you don't.

"You should add them all" is just silly; there's not enough people to
implement them, and not enough screen real estate, and not enough QA,
and so on.

"You should remove them all" is also silly; there are valid reasons
for prefs.

So you need to defend some drawing of the line, something more
practical than "I know it when I see it."

> Well, without allowing the user to make their environment work the way
> THEY work, you will not succeed in your purpose.  Remember, UI stands
> for User Interface, not Developer Interface.  If you want people to
> use it, you have to let them do what they want.  And that means
> preferences.

Aha! This is where I think many traditional Linux users break with the
bulk of users. "What they want" for most users means to write a
document, draw a picture, send email - "what they want" for
traditional Linux users means to play around with the UI settings in
between using Emacs and xterm... ;-)

To say that "what they want... means preferences" I think misses why
most people use computers. But hey, I can't cite the conclusive
200,000-person random sample study, and neither can anyone else.

> I hope you take my comments to heart.  In the mean time, I'm going to
> go look at KDE.  Cheers.

Here is how I see it: the equivalent of replacing the desktop shell on
Windows, or using TweakUI, is to use KDE instead. Or to replace a
single desktop component, such as the window manager, or the web
browser, or terminal emulator. Or to use a TweakUI-equivalent
application with GNOME. That's why it's open source, that's why we use
standard protocols in between the components. You can tweak it until
you turn blue, swap out any piece you want. We aren't going to be

So why have 10 desktop shells and 37 window managers, and have them
all implement all the preferences so they can all be configured to be
just like any of the others? There's no real diversity or choice in
that, just bloat and duplication of effort and oceans of bugs.

With GNOME we're going the try-to-make-it-clean-and-simple route, and
KDE so far is sticking to the more complex route, and ROX and XFCE and
Enlightenment (and Windows and Mac) are trying all sorts of other
variations along the spectrum; and the ones that people like will come
out ahead.  Or maybe several will exist indefinitely because there are
different kinds of people in the world.

If no one uses GNOME, we were wrong; no big deal, other desktops will
soldier on. If only 30% use GNOME, then we met their needs, that's
great too.

But there's no real user choice if all the available software is huge,
bloated, superset-of-everything-ever, hyper-tweakable. Some people
might want to choose It Just Works.

This theory is hardly something GNOME made up:

... etc. At least it's a mass delusion.

But the short answer, ignoring all of the above: I don't think we
should do anything other than what we think is right, based on all the
information we can get.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]