The dangerous myth of the "completely intuitive product"

[Steve Owens <uso01%eagle -at- UNIDATA -dot- COM>]

Larry Kunz (ldkunz -at- vnet -dot- ibm -dot- com) says:
> When I rent a car, or test-drive one at the dealer, I'm glad I don't
> have to read the owner's manual first.  I appreciate knowing that the
> pedals will do what I expect them to.  And I appreciate being able to
> work the gearshift with, at most, documentation that's minimal and
> (literally) at my fingertips:  the P-R-N-D-2-1 on an automatic
> transmission, or the 'H' diagram on a stick shift.

The diagram on the stickshift or the labels on the automatic transmission
are not minimal documentation, they're quick reference cards. The
documentation doesn't exist, which is why you were tutored in how to drive
a car way, waaaaay back when, remember?

> When I lift the hood, I expect to be able to do simple maintenance --
> change the oil, add coolant, replace the air filter -- without consulting
> a manual.  For more complex things, I don't mind using a manual.  But for
> day-in, day-out tasks, I should never need documentation.  Not the first
> time, not ever.

So the first time you ever changed the oil - or changed a tire -
nobody told you what to do?  You'd never seen it done?  You didn't
look for a diagram under the hood or in the owner's manual that showed
you where the dipstick was?

How did you know that the oil even *needed* to be changed?

The answer is that somebody told you.  Whether you learn it from personal
instruction or from a manual, automobile operation and maintenance is
most definitely NOT "intuitive".

The only advantages automobile interfaces have is that they're largely
uniform, so once you know one interface you pretty much know another
interface for an automobile of that class. Go try driving an 18 wheeler,
or a backhoe, or a combine harvester.

This is the same advantages MacIntosh packages seem to have, by the way;
there is a very definite interface standard built into the hardware, and
the predominance of the early applications (many of which came from the
same source, and hence had the same interface) built an expectation, a
standard for what a Mac application interface should be.

The use of intuitive design is like the use of diagrams and pictures.  A
picture can be worth a thousand words in the right situation, but when you
need the words, you need the words.  Intuitive design CAN greatly reduce
the learning curve, but it's only one more tool in the creation of a good
product.  It is NOT a panacea.

Steven J. Owens
uso01 -at- unidata -dot- com
"A program should follow the rule of least surprise.  That is, it should
 do that which will least surprise the user." - paraphrased from The Tao
 of Programming
"The paradigm of a user interface should create expectations in the
 user's mind - and then meet those expectations." - myself