The Psychology of Everyday Things uses several everyday life examples
to show design problems. He discusses problems where people have new
technologies yet cannot use all of the features. He defines several terms and
gives scenarios that relate to those terms.
The visibility of an object is good when the correct parts are
visible and they convey the correct message. Well designed objects should be
easy to interpret and understand. They should have visible clues to help with
operation. Visibility indicates the mapping between intended actions and
actual operations. The visibility of the effects of an action are also
important. A user should get some feedback after completing an action. He
tells the following stories to show his point:
He discusses a set of doors in a post office in a European city that neither
gave any indication whether one should push or pull on the doors nor was it
intuitive which side of the door one should push on. Unfortunately, his
friend was stuck between the doors until he saw someone else go through the
corridor.
He discusses the new phone system at Basic Books that seemed to be disliked by
everyone. Everyone complained there was not a "hold" function. There was a
similar function but noone could figure out how to use it. The new phone
system had poor instructions. It failed to relate the new functions to the
similarly named functions that people already knew about and there was no
visibility (no red button to light up when someone was on hold) of the system.
The affordance of an object are its perceived and actual properties.
Affordance primarily is concerned with the fundamental properties of an
object that determine how it could be used. Affordances give clues
of operation to the user. For example buttons should be pushed and
knobs should be turned. Simple things should not require any explanations.
The design has failed when simple things need labels, pictures, or
instructions.
People often form a conceptual model of a device and mentally
simulate its operation. Conceptual models allow us to predict the
effects of an action. This can only be done when the parts are visible
and the implications are clear. People get clues to how things work by
their constraints, mappings and affordances. He gives examples to
support this: a pair of scissors and a three wheeled bike. The scissors
has holes. The holes are used as affordances (you need to put something in
the holes.) The size of the holes is a constraint so that a limited number
of fingers could fit in them. The mapping (possible actions) between
the holes and the fingers is suggested by the holes. The bike obviously
would not work by looking at the two sets of handle bars connected to the
same wheel.
The operation of an object becomes unclear when there is a false conceptual
model for the system. The refridgerator / freezer temperature control
device showed this. The manufacturer's conceptual model gave the
impression that there were two controls for the temperature of the freezer
and the refridegerator. However, this was not true and made setting the
temperature impossible.
Norton continues to discuss the need of visibility with another phone
system example. The telephone system would allow the user to keep dialing
a busy number until the call could go through. If this process took
several hours the phone could ring and both people could answer the phone
asking who's calling with the belief that the other person placed the call.
The phone did not look complicated but it was still unusable. The mappings
for the functions were arbitrary. Part of the problem was also due to
vestigial features while designing the faceplate for the phone.
Vestigial features are features in a new version of a product that existed
in a previous version of the product. As long as the unintuitive features
of the phone had not gotten any complaints, the features were never
eliminated, and the interfaces for simple things become more complex.
Good mappings and and natural relationships between the controls and the
things controlled make functionality easier to lear and use. Norman discusses a
car system that has over a hundred functions but is much easier to learn
then the aforementioned phone system because the user's expectations, the
necessary actions and the results are meaningful. The placement of the
controls were related to their function and the number of controls was
almost at a one-to-one mapping with the number of functions.
Mapping is refers to the relationship of two things. Natural
mappings refer to cultural standards or physical analogies. These
lead to immediate understanding. For example, the length of a line could
indicate a volume of music. As the volume is increased incrementally, the
line should increase incrementally.
Feedback sending back the user information about what action has
actuall has been done. Norman gives another telephone example to
show his point. When telephones were first invented the user would
push a button and get hear a tone. When the call was being
connected the phone made noised so the user would know what was
happening. Problems with modern telephone systems occur because there are
more features and less feedback.
Unfortunately, it takes about five or six times to get a product right.
If a design for an object is bad it can be considered a failure
after the second or third time. After this the idea is usually dead and
potentially good products are not developed further. The design of
technology usually follows a U curve. At first the product is difficult to
understand. The complexity is decreased to a comfortable level. Then more
features are added and the complexity rises again. The same technology
whose functionality makes life easier, makes life more diffuculty by
making the object harder to use and learn. This is the paradox of
technology. Although the added complexity and difficulty can not be
avoided when functions are added, but this should be minimized with a good
design. It should not be an excuse for a poor design.
I really like all of the practical examples. I have encountered just about all of the systems that he said were problematic.