[UFO Chicago] [mailman-owner@flynn.zork.net: ufo subscription

[Larry Garfield]

Jordan Bettis wrote:

> Reading comprehension is necessary for just about every facet of
> modern life, which is why I suggest that you run, don't walk, to
> your nearest bookstore or library and get this book:
> 
> <http://www.amazon.com/exec/obidos/ASIN/157110089X/qid=1025481289/sr=1-2/ref=sr_1_2/102-9489032-3272912>

Now why is a GNU Fanatic recommending I buy a book at Amazon?  Did the
boycott end without anyone telling me? :-)

> > > "How do I type 'for i in *.dvi do xdvi $i done' in a GUI?"
> > >           -- Anonymous
> >
> 
> > Well that one's easy. Sort your view by file type, drag the *.dvi
> > files and drop them on the viewer app. Next please.
> 
> Of course, it'd be trivial to think up counter examples, what if
> you were trying to recurse over a number of sub directories? I
> just pulled that one out of my ass just now. The point is that
> a GUI could never be as powerful as a good notation. Someone
> who knows the notation will always be more productive than the
> person who is trying to use the GUI.

If you define "powerful" in such a way as to optimize it for where a
CLI's strengths are, then of course the CLI will appear to be the most
powerful.  You are also begging the question.

You are completely ignoring another factor of "powerful"; does it for
me.  If two programs can both produce the same desired output, then the
one that does it with less effort on my part is more powerful, because
it is doing more of the work and I am doing less.  If a decently written
Find dialog can let me recurse subdirectories to find files with a given
extension (hint: many to most can), without me needing to remember what
the syntax for the bash "for" statement is, then it is a quite powerful
program.  It is doing more with less input.  How is that not powerful?

A CLI works very well for certain tasks.  A GUI works very well for
certain tasks.  You can accomplish almost any task with either a CLI or
a GUI of sufficient complexity.  Arguing that one is intrinsically more
powerful than another in all cases goes beyond naiveté to
self-delusional fanaticism.

I do, however, submit one constant: A reasonably well-designed GUI, in
general, will have an easier learning curve than a reasonably
well-designed CLI.  GUIs are more novice user friendly, while CLIs are
more expert user friendly.*  Most people in the world are novice users
of any given program.  

* Usability 101 (aka HCI 310, Introduction to HCI, for the DePaul people
on this list).

<snip>

> As a consequence, the pilot has an intimate knowledge of the operation
> of the craft. He knows exactly what will happen when he manipulates
> those controls, both in the aircraft and in the surrounding
> environment.  When something changes, he knows how to compensate. When
> something breaks, he knows what his options are.
> 
> Now, we don't breed people to be helicopter pilots, genetically
> engineering the to have innate knowledge of the operation of the craft.
> People learn how to fly them by going through hundreds of hours of
> ground school and instructed flight sessions. There is no intuitiveness
> to it. The pilot is the way he is because he took the time to actually
> learn about his aircraft.

We don't breed people to be car drivers, either.  See my earlier analogy
on that subject.  One could easily design a car interface that allowed
the driver to have independent control over all four wheels, each of
which was independently powered and had its own turning machinery.  Such
a car would have far better maneuverability than normal cars.  Heck, the
thing could go sideways!  And it would have at least six controls (one
for each wheel, one for gas, one for break), more if you threw in manual
transmission.  (What kind of luser doesn't want to learn how to handle
gear shifting properly?)  

Of course, we don't do that, because it is better for all involved if
the learning curve for a car is lower.  Turn the unified steering wheel
left, and the car goes left.  (Intuitive!)  Turn it right, and it goes
right.  (Intuitive!)  Push the gas pedal farther to get more power
(Intuitive!).  More people are able to drive, more people can get things
done, more people benefit by the invention of the car.

(Ever wonder why "forward" on a joystick means "go down"?  Because the
air force determined experimentally that forward=down/backward=up was
more intuitive for its pilots than the reverse.  So all planes today are
built with that interface, even though once upon a time they weren't.)

"Intuitive" is a very gray concept, but it is a valid point of
argument.  People naturally think in certain ways.  Design the system
around the human whenever possible, not the other way around.  "Damn the
human, let him learn!" is a great way to make sure that your product is
not widely used.  (See also: GNU/Linux)

> Computers run the world today. The Net is not a toy for academics
> anymore, we can no longer afford to have people who don't know what
> they fuck they are doing trying to maintain systems they don't
> comprehend. That's reality, and if it hurts your little feelings,
> deal.

Computers run the world today.  The Net is not a toy for geeks anymore,
we can no longer afford to have systems that require that the user know
what the fuck they are doing trying to maintain systems they do not have
the time to comprehend.  That's reality, and if it hurts your little
feelings, deal.

Do you consider the Sun/Cobalt Qube to be bad because it offers turn-key
power to people who don't recompile their kernels for fun?

> > Like I said before, you are such a condescending prick.
> 
> Everyone needs a hobby.

I was not aware that your hobby consisted of actively working to prevent
the philosophy of Free Software from succeeding by insisting that it
remain completely useless to 99% of the population.  I'm sure Richard
Stallman would be proud.  (And I say that with no sarcasm in the
slightest.)

-- 
Larry Garfield			AIM: LOLG42
lgarfiel@students.depaul.edu	ICQ: 6817012

-- "If at first you don't succeed, skydiving isn't for you." :-)