[Usability] Re: [Desktop_architects] Printing dialog and GNOME

[Havoc Pennington <hp pobox com>]

Hi,

No clue what the history of PPD stuff in GNOME in particular is, but
if I can go on a little tangent about the best approach to sorting it
out with the usability people (or rather, software designers).

The "XYZ is too confusing for idiots" line is kind of a caricature
strawman of a position, stated only so it can be shot down.  No good
software designer is thinking about their work in the way that
engineers and sysadmins seem to think when they say the "confusing to
idiots" line.

 - "avoid being confusing" isn't the goal; the goal is to provide
specific functionality/experience
    to some specific (hopefully well-understood, maybe narrow, maybe
broad) set of people.
 - "idiots" hopefully isn't how one thinks of this set of people, the
closest accurate
    words might be "busy" and "not interested in our software for its own sake"
 - "universal design" that's good for as many people as possible is
one ideal, though
   not always possible

When designers try to avoid clutter I think they'd consider it more
"distracting" or "not the most important use of limited engineering
time and resources" or "evidence of a root problem with the design" or
"lack of focus" rather than "confusing" (in my experience).

Think about car design; what's clutter on an SUV is not the same as
what's clutter on a Ferrari. Cars are not designed in order to "avoid
being confusing to idiots," but neither is every car a combination
pickup/station-wagon/SUV/sportscar that's all things to all people.
There are cars that I love and that I hate, but (almost) every car has
someone that loves it.
Of the cars that I hate, I'm not really "confused" by them, but I do
find that they don't appeal to my lifestyle or tastes.

If you approach an interaction designer or usability team with an
argument like "there should be no tradeoffs, just dump it all in!
nope, no clue what the use-case or root problem is! yee-haw!" they'll
respond about the same way most software maintainers would if you
submitted a patch with that rationale, or about how a car designer
would if you proposed the universal car for every situation and
person.

I'm just suggesting that if you wanted to talk to people about this
it'd be good to cook up some details. What kinds of things can
printers do? What kinds of users use those things or buy those
printers? How do those things fit in to the person's work or fun?
Designers aren't going to want to talk about the idea of "stuff in the
PPD file," they're going to want to talk about "stuff people want to
do."

I think most designers would be primarily worried that something
important in the PPD stuff is only exposed as autogenerated cruft UI
on the advanced tab. i.e. they'd want to be sure anything important
had a good, visible UI somewhere (in the dialog or somewhere better).
Once a designer understood the set of important things, I bet they'd
be fairly indifferent about dumping all the rest that a couple people
somewhere possibly care about on an Advanced tab.

BTW people who are supporters of design decisions often _also_
misunderstand the point as "not being confusing to dumb users" and
advocate the changes under that formulation. Which is kind of
counterproductive and fails to emphasize what's genuinely worth
talking about. I'm sure you can dig up some quotes from me in the past
where I didn't get this either.

</intended-to-be-helpful reasonable-sounding comments>
<personal notes>

Off topic even further into personal editorial, I don't really agree
with Jeff's view that everyone can be happy. I have the view that you
can't optimize for everyone and that there are two kinds of products:
    a) the kind that lots of people love and lots of people hate
    b) the kind that everyone can live with
My personality leans a bit toward a) and I can even get a perverse
kick out of making the right kind of people hate something ;-) but b)
is probably more right for a "commoditization" product strategy and
the adults in the room.

Cute blog post on this, with pictures:
http://headrush.typepad.com/creating_passionate_users/2005/01/be_brave.html

The reaction Linus has probably comes in part from when I was having a
good time actively antagonizing people like Linus rather than just
quietly focusing on different audiences. His followup mail mentioning
the window manager seems to support my guess that it's my fault ;-)

Given finite resources, I think it was more often than not correct to
make the decisions that Linus and friends complain about, and
(personally) I think it makes GNOME a better desktop that we did make
them. I'll skip the reasons why here, they are mostly historical
accident anyway.

At the same time, it wasn't really necessary for me (or whoever else,
don't want to be egotistical here) to troll while doing it. I'm glad
people like Jeff don't have my little quirks ;-)
and he's much more representative of today's GNOME (which I have no
influence over to speak of).

I'm typing this in a gmail text box that is maybe 15 lines high,
clearly optimized for people who send reasonable-length email instead
of Havoc-length email. I'm sure reporting this to the gmail people
would result in them thinking "we aren't going to optimize our stuff
for this crazy nutjob" but they'd probably be clever enough not to
explain it that way which is where they're one-up on my window manager
maintenance techniques.

I do worry that while in the past both GNOME and KDE were firmly in
camp a)  - with different audiences and thus lovers/haters, making the
"two desktop" thing halfway useful even - more recently they are both
moving toward b). But I trust the current leadership of those projects
to figure out the right path and walk this fine line nicely.

Havoc