Re: Arlo, a little QA comment regarding your interview with linux.com

[John Sullivan <sullivan eazel com>]

"Guillermo S. Romero / Familia Romero" wrote:

> sullivan eazel com (2000-10-25 at 0802.24 -0700):
> [...]
> > I understand your point, which is that some "preferences" are really just
> > features that aren't used much, and are thus relegated away to a preferences
> > dialog rather than being put into a more obvious place in the program.
>
> What about options that user never use cos never think about them or
> cos they are so hidden than user preffer keep in one mode, but would
> use more if they know about them / were easy to change? How can that
> options be detected, so the can be put in a better place, and docs be
> writen explaining what the do?

Part of the answer here is to design the set of features so that things are
clearly labelled and logically grouped, and thus nothing is too difficult to
find.

Part of the answer here is to avoid having so many preferences that important
ones get lost in the noise of the trivial, unimportant ones.

Part of the answer here is to carefully consider each potential preference and
trade off its visibility vs. its importance. Not everything can be equally
visible (or rather, if everything is equally visible, then everything is equally
hard to find too, and for a large, complex program, this is clearly
unacceptable).

Part of the answer here is to determine through usability testing and other forms
of feedback which preferences are more important. Exactly how to do that is an
art, not a science, and certainly does not have simple well-understood rules that
apply to every situation. I'm pretty sure democratic vote will not result in the
best solution, though it might be valuable input.

John