[Arcana] More Emacs love.

Karl Fogel kfogel at red-bean.com
Sat Jul 28 14:32:34 CDT 2012

Jim Blandy <jimb at red-bean.com> writes:
>How can a feature be discoverable if the user's eye needn't move at
>all, but the feature doesn't cover anything?

I'm not sure I understand the emphasis on discoverability?  All the
interfaces being discussed don't show anything until the user invokes
them.  You have to already know where to find it before it even shows

There's the question of discovering *how* to use it fully, once you've
invoked it -- is that what you're getting at?

>> If so, the answer is that I think there's an inevitable tradeoff between
>> interfaces whose features are discoverable (i.e., friendly to newcomers)
>> and interfaces whose features do not distract when not invoked (i.e.,
>> friendly to long-time users who make an investment).
>> A UI can be easy to use when first encountered, or it can reward
>> investment.  But I don't think it can consistently do both.  At least,
>> the two ways will come into conflict pretty often.  And I'm not saying
>> one way is better than the other, in case that wasn't obvious.
>> Immature fields talk about "good" vs "bad".  Mature fields talk about
>> tradeoffs.  It has ever been thus.
>I would temper this with the acknowledgement that making tradeoffs is
>what everyone does at the limit of their knowledge. We're not terribly
>experienced UI designers here.

Couple of points:

Are you sure that's true?  We're very experienced with a particular kind
of UI.  There is a class of UI designers who concentrate on a different
style of UI (newbie-oriented UIs) who sometimes mistakenly think that
"newbie-friendly" and "good" are synonyms.  Those people are usually
horrified when they see how Emacs works.  Since people who use Emacs as
it needs to be used -- treating it as an "investment user interface" --
have a pretty good experience with it, I'm skeptical of any UI designer
who doesn't understand the tradeoffs behind that fact.  Which is a lot
of UI designers, in my experience.

And I'm not sure that tradeoffs are a sign of being at the limit of
one's knowledge.  Maybe they are, but I actually would have thought the
opposite.  What I was trying to say is that tradeoffs are inherent in UI
design as in just any field, so the better you know that field, the more
likely you are to being making tradeoffs on purpose.

I keep trying out the UIs that such newbie-oriented designers create.
Sometimes I try them for a good long time, taking care to explore the UI
and really learn it.  And yet, often by the end I'm still left convinced
that my use case would be better served with an Emacs mode (or the
equivalent in that environment -- that is, a keyboard-centric text
interface with an extension/customization language).

It doesn't mean those UI designers are wrong, per se.  It might just
mean people like me are too small a market to serve.  But if they're not
at least making that tradeoff on purpose, then I think they're missing a
crucial thing about UI design.

>Perhaps someone could make the UI appear *behind* the text, in a
>not-too-saturated, contrasting color. Perhaps, if it were used
>consistently, the color alone would suggest which modifier key to
>start with ("When a Meta- command could apply, we always mark its
>target with green..."). I don't know. It just seems kind of arrogant
>for this band of merry travelers to declare that they've exhausted
>their options.

I think we've had this conversation before, and that basically you
disagree with my contention that a given UI can never be simultaneously
good in all the ways it is possible for a UI to be good, because some
ways are in direct conflict with other ways.

I certainly wouldn't declare that we'd exhausted the options.  But I do
think it's pretty obvious in the UI in question that there's a tradeoff
between visual motion and time motion: you can present all the options,
but then the eye has move to scan them; or you can have successive
keystrokes cycle through the options (from most likely to least likely),
thus allowing the eye to stay focused on the current choice.  Which one
is better depends a lot on the user, and some on the number of choices,
the context, how you present the choices if you choose to do that, etc.
That's why I didn't say flyspell's way was worse -- I merely said it was
worse for people like me :-).

More information about the Arcana mailing list