🚀 go-pugleaf

In article <66.66.0.7.2159dd2e@infos.msk.su>, dat@infos.msk.su (David Tolpin) writes:
>In a msg on <Sep 27 14:27>, smith@ctron.com  of 66:66/0.999  writes:

> > Ulrich may be right that it can be learned, the main point from a 
> > user-interface perspective is: it shouldn't have to be.  

>>From this point of view the worst examples  of user interface is a pencil,
>very long time must be spent before one starts to use it intuitively.

Not so.  A pencil's operation is intuitively obvious, and even a baby can begin
making marks with it very quickly.  Once one has learned to use a pencil to
draw or make letters, the knowledge generalizes easily to crayons, brushes and
paint, charcoal, etc. etc.  After all, it is based on prior art: drawing one
stone across another.

> > There are 
> > two reasons. Firstly, it is non-intuitive.  The only similar idea 
> > to be found outside of computer science is "chording" in music.  
> > Few people find that an easy idea to handle (and the exceptions 
> > merely prove the rule).  

>Several examples of doing two actions concurrently:

>car driving,
>tailoring,
>most of arts, including pictures, sculptures etc,
>mixing milk and tea at breakfast,
>and many other things.

This is a counter-example?  A typical computer-user is _already_ multitasking,
the mouse action is not an end in itself, the user is trying to _accomplish_
something.  Unless the mouse action can be made below the conscious level, it
will interfere with the process, and none of your examples require chording.
Of the examples given that would require similar digital dexterity, both art
and tailoring require years of training and practice.

But moving a steering wheel is intuitive, so is pushing a brush, so is using
a needle and thread, pouring milk and many other things.  Each is directly
linked to a real-world paradigm.  Chording is not, except for music, which
is an activity that takes more time and effort to learn than using a mouse-
driven computer application.  It is therefore silly to use paradigms from it
in order to make computers easier to use.

>And difference between pressing,clicking and double-clicking is like staccato. 
>That is not easier than chording.

The difficulty of double-clicking is why the Motif and Apple both recommend
that it be reserved for activating _default_ actions.  People who can't or
don't wish to use double-clicks may simply select with one click and then
pull down a menu.  In the Mac Finder, click to select, then choose "Open" from
the File menu.

But for the record, and I think Apple's human factors people will back me up
on this, double-clicking is much easier than chording.  For many people, it's
actually easier than trying to remember what the other mouse buttons are for
in the first place, which is why Apple went with a one-button mouse in the
first place.  Double-clicking is not an integral feature of the design, it's
for experts.

> > Secondly, it presumes a degree of manual dexterity that many people may 
> > not have.  

>Why? I didn't read about the evidence (at least patterns desciption) that many 
>people cannot move their fingers separately.

Moving fingers separately is not the issue, it's moving them together in close
enough synchronization for the chord to be detectable by the system.  Flexing
one finger muscle to double-click is quite trivial by comparison.  In any
event, Apple spent millions on basic human-factor research when they first
evaluated mouse-driven interfaces.  If you will remember, the inventor of the
mouse (whose name, I am distressed to admit, will not come to mind) invented
it to go with a small piano-style keyboard - it even had white and recessed
black keys like a piano - which was to be used to chord hundreds of different
commands.  It was fast, efficient, and extraordinarily effective.  It also was
just as complicated as learning to play a piano.  Most people can't play a 
piano.

Your own facility with such movements is to be congratulated, but do not
mistake what is easy or good for you with what is easy or good to learn or
use for people to whom computers are just business appliances.  Human factors
is a science in its own right to train people to make that differentiation.
I have used mice for years, I am familiar with most human/computer interface
paradigms that have been tried, and I am trained in human factors and user
interface standards.  I know how hard it is for me to judge and reject a
feature I like and would use because it is a poor choice for most users, and
I know that most computer types just can't make that distinction.  That's
why we still get systems like Oberon.  I can use Oberon.  But I'm an expert,
I can use anything.  But I've dealt with many people who hate Macs and Windows
because the hand-eye coordination involved is already borderline for them.

> > It used to be people objected to the mouse on the grounds of having to 
> > remove their hands from the keyboard.  These complaints, in my 
> > experience, have dropped off dramatically (though they are still heard now 
> > and again), but have been replaced by what amounts to complaints about 
> > an over-reliance on hand-eye coordination.  

>It is not over-reliance, the mode life requires often much more sofisticated 
>things to do.

If that were true then all the human-factors people in the world are completely
wrong about how and why people interact well with computers.  I've seen the
studies done by IBM, Apple and the US Gov't.  I'm waiting to see yours.

> > In short, many people 
> > just feel clumsy with the mouse, but are confident of their ability 
> > to eventually hunt-and-peck even the obtuse command.  To add to the 
> > mouse a sophisticated repertoire of two and three finger operations 
> > compounds this problem.

>I don't think that presshing Ctrl-Shift-F12 to print a document is easier than 
>to press two buttons on the mouse. By the way, don't you think it is wise to 
>make cancelling a command a bit more difficult than invoking (in Oberon 
>cancelling is the only case of using three-buttons combination.)

Ctrl-Shift-F12 is not easier than pressing two buttons on a mouse.  But I do
maintain that pulling down a menu and hitting "Print" is easier yet.  You're
giving me a choice between bad and worse, when good has already been defined.
The world is not just Windows/DOS.

In most graphical UI systems, cancelling is done by single-clicking the
"Cancel" button in the dialog box, which I find much easier and more
mnemonic than chording a 3-button operation.  And the fact that it is
the "only" three-button operation in Oberon does not make it special or
unique, there are four and five button mice on the market, how is a
3-button chord special or mnemonic for them?

> > The Oberon system's interface is an experts design - like a Unix 
> > shell, it is very powerful but it relies on many non-intuitive 
> > concepts, and so is a big pain to learn.  

>The concepts are non-intuitive for one that already learned other ways of
>using mouse. And the opposite would be if Oberon System apper first. 

Oberon did not appear first.  As a followup, it bears the responsibility of
acknowledging and using prior art, or of presenting something vastly superior.
It has done neither, and so it is doomed to be a quaint little backwater
experiment.  The language itself has potential - I hope it will be evaluated
outside of the system it was used to create.  But I fear the system will be
of only academic interest, and then only to academics.

> > It is, in fact, built on 
> > a weird, mouse-driven command line paradigm, combining the worst 
> > features of both.  

>Command line paradigm is not the same as calling command with parameters.Which 
>worst features of those two paradigm are inherited by Oberon? And which best 
>are forgotten?

Isn't it obvious?  The entire concept of a command line.  That it can be
built with mouse clicks is merely window-dressing, the system must still
construct it, limit itself to things that can be expressed with it, and
waste time processing it.  There is ample research showing that presenting
a manipulation paradigm is clearly superior.

> > It requires a new user to internalize a 
> > considerable number of concepts and "finger macros" - far more than 
> > we now know is needed.  None of this would be objectionable if 
> > this interface was even a little bit standard - but nothing else is 
> > even remotely similar, Wirth ignored virtually _all_ prior art but 
> > his own.  

>Well, actually this is the heart of your article. The most significant problem 
>with Oberon is not that it is built upon one or another paradigm, and is not 
>that double-clicking is easier than pressing two buttons, but just that Oberon 
>does not follow accepted traditions in user-interface.

It must follow them or better them.  It has done neither.

> > You cannot use Mac or Open Look or Motif concepts in the 
> > Oberon system - nor can you use anything you learn in Oberon in any 
> > of the others.  From a simple cost- effectiveness point of view, 
> > Oberon looses.

>From a simple cost-effectiveness point of view Oberon looses, as well as 
>everything that isn't made looking at the simple cost-effectiveness. What is 
>done is that an alternative approach is offered. And it is sientifically 
>incorrect to judge it from the point of view of Motif and OpenLook trained
>user. 

I'm not sure just what "science" you refer to with _that_ remark.  The Mac
interface, Motif and Open Look were firmly based on extensive human factors
research, Oberon's was not.  Oberon is not clearly and unequivocally superior
to the average user, in fact it has a ferocious learning curve.  If one climbs
that curve, does one experience greater productivity, have they become better
workers?  No, they can't do anything a Motif or Apple or Open Look person can't
do, and do more easily and frequently just as quickly.

There is no better way to evaluate anything than cost-effectiveness.  What is
the point of doing something less effective and/or more costly?  Research?
Well, that research is already being done, and it shows that Oberon and its
chording is an interesting little historical byway.  It's too bad that that
was already pretty well-known to most human factor experts before Wirth even
started.

No one evaluates new interface designs from the point of view of a Mac user,
or Motif user or Open Look "trained" user.  We look at it from the point of
view of someone sitting down to teach a new user how to use a computer.  If
you put Oberon, Apple, Motif and Open Look-based systems in a room, with
instructors, and set down new users at each, the Apple user will be doing
useful work first, with the Motif and Open Look-based people coming right along
behind, neck-and-neck.  The Oberon (l)user will still be thrashing about trying
to figure out what you are talking about long after the others have finished
their starting projects and gone home.  I know this because I've read up on
Apple's research, they tried just those sorts of experiments.

Once this has been accomplished we then look at the so-called "power" user or
"expert" user.  Each system must present ways of accomplishing things more
quickly when someone "knows what he is doing."  Oberon competes better here,
but it simply is not better.  When experts are even, the learning curve
determines the winner.

As an interface for experts willing to pay the price and willing to forego the
advantage of learning a standard interface, Oberon has some value.  As a
commercially-viable alternative to Motif or the others, Oberon comes up years
late and many dollars short.  Wirth just didn't do his homework on that one.

Not that I expect commercial realities from Wirth.  I used to, but I decided
I was expecting too much.

Larry Smith (smith@ctron.com)  No, I don't speak for Cabletron.  Need you ask?
-
How about we all do something reasonable?  I bet that'll confuse 'em!