Tools design (was Re: Metacard support)

[Richard Gaskin]

jbv wrote:

> Well, I'm not sure... depends on what you mean by
> "customizable"...
> ...the (apparent) ease of use of HTML and Flash has lead
> to great productions, but also (IMHO) 100 times more crap.

And with the recent browser plug-in lawsuit most of it is now inconvenient
as well.  :)

> The birth of the Mac interface in the mid 80's had lead to
> some very strong & efficient interface design set of rules,
> (for instance having a File menu and an Edit menu with more
> or less the same items in each app) that was indeed a very
> good thing after years of anarchy in the Windows environment.
> But the widespread use of HTML and the many other development
> tools have started to dilute those rules (many newbies aren't even
> aware of them). You have no idea what I've been asked to do by
> clients with no computer interface culture... It's sometimes hard
> to resist...

I'm of two minds on this: While I regularly cite HIG verse and chapter, I've
lately become enamored of Eovia's Carrara, a truly unique UI that IMNSHO
embodies most of the underlying principles of good design while completely
blowing off nearly all of the specifics suggested by OS vendors.

I haven't yet defined for myself the principles which would tell me when
it's appropriate to consider such a radical departure from the norm, but I
think part of it has to do with how long a session with the app lasts (i.e.,
whether other apps or the Finder are useful or distractions), and whether
the nature of the tasks suggests a more isolated workflow is beneficial.

In our HyperRESEARCH application we got a lot of requests for a backdrop.
Since that's usually used for multimedia presentations I initially resisted
adding it to a productivity tool like HR.  But on further consideration I
recognized that a large part of the workflow involved many days, sometimes
weeks, of using one set of HR features in a repetitive workflow for six to
ten hours a day (such is the nature of qualitative research that some of the
most time-consuming tasks require subjective human judgement).  In such a
workflow anything else happening on the machine is a distraction, and the
backdrop helps people focus on the task at hand with far less visual "noise"
in their computing environment.

Unlike Carrara, HR's backdrop is optional, which leads us to the importance
of choice implied here:

> Well, as long as end users only design crap interface for their
> own IDE, it's not a too big problem (although it might affect
> their productivity by several orders of magnitude)...

...and in some cases positively.  Geoff Canyon's Navigator is very powerful,
but like Emacs it's not as intuitive as less powerful alternatives.

In many cases there is a trade-off between learnability and usability.  A
bicycle doesn't require weeks of specialized training to use, but can't take
you as far as a car (though a car doesn't provide the energy efficiency or
health benefits of running errands on a bicycle equipped with panniers, but
that's a thread for alt.energy.homepower).

> But their choices might extend to apps for clients, or stuff they
> put online for others to download...

Software success is a democratic process:  crap doesn't get used often, but
even unusual interfaces liked Geoff's Navigator or Emacs can become very
popular if they deliver the goods on productivity.

> I have the feeling that switchable (on/off) parts of the interface
> (palettes, etc) might be a better option...

Bingo - it's all about choice.  Mix and match as you like.

While there's a good argument to follow HIGs in light of the benefits of
consistency between tools, remember that one of the sad facts of modern HIGs
is that they are often driven by factors other than usability research.

For example, Microsoft has one of the most well-funded UI research labs in
history, yet all of their UIs ignore Fitts' Law (see
<http://www.asktog.com/columns/022DesignedToGiveFitts.html>), placing the
menu bar below the drag region of windows rather than at the top of the
monitor which provides a natural backstop for the pointer.  They're not
ignorant of Fitts' Law, they just have to distinguish themselves from Mac OS
due to legal limitations.

And as for Apple, although the OS design was historically driven by UI
research, Gil Amelio fired the goose that laid the golden egg as a
cost-cutting move, and when Jobs returned rather than hire them back he
hired mostly graphic artists instead, making a UI that looks great and demos
well but has a number of attractive-but-counterproductive "features" (note
that transparency of menus and window trimmings made them almost unreadable
in 10.0; to Apple's credit they do seem responsive to user outcry; for more
see <http://www.asktog.com/columns/034OSX-FirstLook.html>).

Where HIGs conflict, one would assume that one of the vendors has better UI
research.  While that may have been true back in the days when Tog regularly
published Apple's research methodologies and results, today very little UI
research is presented by OS vendors to substantiate design decisions,
raising questions of which decesions are the result of research and which
are mere edicts.

Meanwhile, multi-platform developers sit in the middle of the OS wars, like
third-world countries trampled in the proxy wars of the Cold War era, trying
to decide if the confirmation button in a dialog should be on the right as
in Mac OS or on the left as in Windows. ::sigh::

Fortunately, with development tools we're only addressing a relatively small
and homogeneous subset of the larger demographic, one more inclined to be
forgiving of inconsistencies between tools as long as the power is
self-evident.

-- 
 Richard Gaskin 
 Fourth World Media Corporation
 ___________________________________________________________
 Ambassador at FourthWorld.com       http://www.FourthWorld.com
 Tel: 323-225-3717                       AIM: FourthWorldInc