Re: UI Guidelines: Dialogs

[Calum Benson <calum benson ireland sun com>]

thristian atdot org wrote:
>
> I'm not sure if this is offtopic or not... *how* would you test such a
> thing? The combinations to be tested are (I guess) Back/Forward,
> Back/Next, Prev/Next, Previous/Next, Prev/Forward, Previous/Forward
> and any others we think of.

It's a perfectly fair question!  For something like this, I wouldn't say
it's really a matter of timing everything with the stopwatch-- as you
said, the time differences would be negligible.  And the purpose, layout
and functionality of druids is such that (provided different druids are
consistent) people should very quickly learn how to use them anyway--
probably even if you wrote absolute nonsense on the back/next buttons,
and  showed them a few times what those 'nonsense buttons' did.

So in this case, I think it's more a case of finding the words that the
users are most comfortable with-- the ones they would use themselves to
describe what they're doing.

One of the options you mentioned would certainly be a possibility--
asking people to explain what they think each button in the dialog
does.  This is  part of the 'thinking out loud' approach that you'd
often use when testing a prototype anyway.  As you said, though, it
would be hard (or at least a lot of work) to find out which labels were
"better" with this approach.  But if you got lucky, you might find that
one or two labels were being consistently described more correctly or
incorrectly than the others.

I'd probably turn this one on its head, though, and ask people to
suggest what *they* think should be written on the buttons, having
explained to them what those buttons do.  (The explanation would have to
be very carefully worded so as not to influence their choice, though!). 
This would really only work with people who hadn't used a wizard/druid
before, but they're probably the people to whom the button labelling
matters most anyway.  (I can think of other ways you could test with
people who *had* used a druid before, but I can feel this email getting
too long already...)

Of course, as with any usability test, there's no guarantee that you'll
get a clear majority decision whichever way you approach it.  The best
you can do is remove as many of the variables as you can, then use the
results and your knowledge of the test particpants, the eventual target
audience and any relevant design guidelines to make the best decision
you can.  Let's face it, if it was easy, Windoze wouldn't still suck
after all these years  :o)

Cheeri,
Calum.

-- 
CALUM BENSON, Usability Engineer       Sun Microsystems Ireland
mailto:calum benson ireland sun com    Desktop Engineering Group
http://www.sun.ie                      +353 1 819 9771

Any opinions are personal and not necessarily those of Sun Microsystems