4

THE HUMAN INTERFACE

Perfect simp/idly is unconsciously audacious.

-George Meredith

4.1 MAN-MACHINE DIALOGUE AND INDUSTRIAL DESIGN

The interaction between a user and an electronic instrument is compJex. Look,

feel, operational flow, perception, and even documentation compose a dialogue
between user and machine. It's more than display layouts. It's more than style and
packaging. It's the entire instrument communicated to the user, from the front
panel to the last nuance of program flow. That dialogue is the human interface of
the product and defines the set of interactions, encompasses the user's view and
understanding of your instrument, and affects the entire design of a product.
The human interface should influence everything that you do during develop-
ment of a product. It is communication, and it establishes the message and the
dialogue. The discipline of industrial design can help you define the message,
establish the man-machine dialogue, design the human interface, and ultimately
design the product. Industrial design is a multidisciplinary process that uses many
concepts to refine the design of the human interface. It involves the following
(Couch, 1993):

1. The design team: engineers, software developers, industrial and graphic de-
signers, model makers, and writers
2. Stakeholders: end users, influencers, customers, and clients
3. Repeated discovery, development, and delivery
4. Early model, concepts

In general, don't just design features; design a product with its whole purpose
and function in mind. Don't fall in love with your ideas; users can have very differ-
ent ideas about what is important. Consequently, you must integrate users into the
design process to understand their needs, wants, and desires.

Scanned with CamS

 User acceptance and satisfaction help measure the.success of an instrument.
According to Marcus and Van Dam (199 I, p. 52), "designers ha~e le.arned that a
good product must be easy 10 produce, market, learn, llS~, and mum.tuln." Further-
more, subjective issues such as "look and feel" and uuracuvc packaging arc equally
important to users and your cllstomcrs. ., . .
This chapter introduces guidelincs nnd pJ'lllclples that drive the ~eslgn of hu-
man interfaces for devices, appliances, and instruments. It starts with the nee<.l~
and desires of the user, goes on to define the clements of design in a human inter-
face, and ends with some specific concerns. YOII should carefully define the inter-
face before code is written, circuits are desigua«, or packages are buill.

4.2 USER-CENTERED DESIGN

Understanding users is essential to successful design. Por instance, McDonald and

Schvaneveldt (1988) cited a study that showed pilots could access the information
from a cockpit display more efficiently when the menu structure was derived from
judgments of the pilots rather than developed by a design team.
Collaboration with the user during design wi II improve the chance of success
in the final product. Including users in the process helps eliminate false assump-
tions and estimations. Consulting with users during design also encourages them
to "own" the design solution and enhance their acceptance of the instrument

Example 4.2.1 Don't take a solution or some interesting technology

and go looking for a problem. A former officer in the U.S. Coast Guard
explained that when programmable calculators became widely available
a number of people developed programs for calculating values in celes-
tial navigation. But crews did not use the programs and calculators be-
cause the manual preparation of the charts was the only departure from
the monotony of the cruise and they looked forward to the calculations
each day.

4.2.1 Users, Influencers, and Customers

First and foremost, establish who will use, influence the use of, and buy your
product. You must know the difference between the customer (buyer) and the actual
user. A user is the predominant operator of your instrument. Generally users are
the most familiar with its function, though they are not always knowledgeable
about its operation. The customer is the purchaser of the device. An influencer
exercises considerable control over the purchase and use of a product and may
have considerable knowledge about its operation.
Successful companies meet customer expectations. Site visits, resident
knowledge, user profiles, focus groups, and concept descriptions will help yOU I
~cannea wnn t.;amS
 , 'J J

£UUDpJe 4.2.1.1 Consider a POrtabl .

{ydients. The patient is the user. Th ef.m~dlcal device for ambulatory
&. th • . e amity may c.
re fill S lor e nevrce and therefore ,requently purchase
. acts as a cu t
influencer. Certainly the doctor who . S erne. and possibly an
. h prescnbes the d '. .
with muc control over the use ofth eVICC IS an mfluencer
• _v e product N .
t211iS who tram the patient can also e : urscs and medical ar,sis-
use of the P;UdUCL xerr consIderable influenceOYerthe

c:::d~rsmnd the population of users. customers d.

ties and expectations The follow' h ,~n mfluencers and their capabili-
, . rng t rcc sectrone describe processes anal- ses
sa restsmat define the users customers and . Il . , Y.
_ • '. •. ,In uencers and refine the design
ccacep of a product with their Interests in mind.

~,2_ Prescription and Process

~Yo.:_ceed ~oknow th~~~se for your device before knowing how to implement
::. 1ce desires, capabilities, and expectations of users, customers, and influencers
...~ define the purpose for your device. Don Norman (1986) gives the followins
~.eK::ip:ion for design: -
• roo user-centered design: Start with the needs of the user. From the
cser's point of view, the interface is the system, ... User-centered design
emphasizes that the purpose of the system is to serve the user, 00£ to use
2 specific technology, notto be an elegant piece of programming" (p. 61).

• lake interface design seriously as an independenr and important

problem: It takes at least three kinds of special knowledge to design an
interface: first. knowledge of design, of programming. and of technology;
second .. knowledge of people, of the principles of mental computation.
or communication, and of interaction; and third. expert knowledge of
tee task that is to be accomplished .... A truly user-centered design ...
will have to be done in collaboration with people trained in all these
areas" (p. 60).
Toe prescription states what should be done; th~ following_process t:~how
:: =, done. Applying this process will identify me major elements of the user Inter-
face (Card et al., 1983):

1. Analyze user needs.

2. Specify the following:
a Performance requiremeD{S
b. Tasks
c. ~lethods (0 do the tasks

~-- ... -_._.. ... _ .. -_.nS

102

. th ds for a task to be clear to the user and easy toa

Design alternatlVe me 0 PPly.
3.
4. Design for error ~e.c~veryipredicted performance and compare with assu
Analyze the sensitIvity 0 . mp'
5. . Use prototype and field testmg. .
nons.
DoD Standard 2167 for the U.S. military
t d standar dS sue h as ,
DocumcI.l e. 'bl ms but they can provide only a rough oUtlinefor
1 orne obvious pro e , . f .
he 1p catc 1: f 1 design is not the domain 0 a smart deSIgnerOr
the human \tlterfac~. A ~~ccelss.uit is the synthesis of insights, opinions, andpref.
even a team of designers: rat let, . li h
fi ted parties including the c lent, t e users, andthe
crences of a group 0 mteres , ,
design team.

4.2.3 Analysis
Refining the human interface requires analyzing the t~sk, understanding the users,
and te ring prototypes. Analysis may take the following forms:

• Task analysis
• Questionnaires
• Informal and formal interviews
• Focus groups
• Alpha test sites
• Beta test sites
Begin development of the human interface by performing a task analysis. Get
to know potential users through questionnaires and interviews. Early on you can
use informal sessions to define the interface. Later in development you will want
to formalize the interviews to compile useful statistics.
Focus groups are a good source of primary research to set direction for prod-
uct development. A focus group has a moderator leading a group of potential users
in a free-form discussion of the product's utility. An impartial moderator will have
wide latitude, steering away from the how, and focusing on the what.
Alpha test sites allow close interaction between designer and potential users
to refine designs. Beta test sites require potential users to evaluate the function of
devices in nearly finished form, with little or no interaction with the designer.

Example 4.2.3.1 A company collaborated with an industrial design finn

to develop lunch boxes for school children aged 6 to 12. It extensively
studied children, how they used lunch kits, and concepts for lunch boxes.
The analyses took several routes; they asked children to do the
following:
• Sketch lunch boxes.
• Color in line drawings of lunch box concepts. .. ' , .

Scanned with CamS

 • Place color chips On lunch box conce t
di . p s.
• Tn reate what characters they wanted tl id .
t on ie 51 e ot the lunch box,
The company nlso built models or the ' d
, . I 'I .I '.. . concepts an asked parents
anc C 11 ch l:11 to cvnluate the concepts (N . TI '
.~. I ,. . . , . , otc: le users were chil-
(hen. t ic ClJstOI11Cl s were their parents •(a 11d tlic In
infl
uencers were peers
unc I parcnts.)
" The . researchers
,,' found that. children
. wa n t eu.1 diff
I J erent compartments
1,01, th~1 rnos, ~and\vI,ch, and Irun. Interestingly, they also found that the
child: en had ecological concerns and wantcd cnvl
, • ." 'HI cnvtronrncnta IIy salec
ma-
tcrials, 1 he new lunch kits sold very well (Couch, J 993). (While the
lunch b~x W~lS n.ol an electronic product. the principle of investigating
user desires IS still npplicable.j

Obviously, in\l~lvjng users throughout development is integral to these analy-

scs, but you should follow several maxims when dealing with users. First, strive to
under. tand users: your knowledge will always be incomplete. Second, don't nee-
e. sarily try for consensus; accept and work with differences, Third. you wilJ have
to balance the wants of the user against your risks to develop a minimum product
with maximum benefit.

4,2.4 Rapid Prototvping and Field Testing

Rapid prototyping and field testing are two methods that implement the interaction
and collaboration between designers and users, Rapid prototyping allows the
interactive development of rudimentary functions by the designer and user in an
operational environment. A graphical interface may be simulated with screens that
represent. but do not implement certain functions, thereby speeding development
and isolating problems early. Field testing involves more mature products, takes
longer. and allows users to examine functional products,

Example 4.2.4.1 J drove a car that had a control knob for the dashboard
lights mounted on the upper edge of the dash near the rim of the steering
wheel. When I made left turns, my knuckles would occasionally rake
across the knob and shut off the dash lights even though my hands were
gripping the steering wheel properly. Thorough prototype and field test-
ing would have revealed that the knob location was poor and needed to
be moved.

R aptid prototypmg
. an d fiIe.ld testing are absolutely necessary
~, to the successful
d eve Iopment 0 f anan imstrumen t . Rapid protot)lping and field testing,. can speed
D theid
d eve Iopment process b y revea ling 1
problems and illustrating applications, aVI
 I I I \ ••.•

. S ttl' vnllcy, C~lIiforni;1.... rated. "I'd h,''Ie t

I)F Inc. III • CO ., • f 0
~hlliC(1I1i pre, idcnt of I '-1 Ily hnq mort.: impncl on a success ul program Ihan
• 'flg I)rol)rt ) ay ..The act I
say that the prO~?IYP~" 90,). 19(». He went on to sav, /I Il~ usen, I"
'lilY other phnsc (rr.lser. J 9 I .itcjeedb(lckjro/ll what we expected (p. 198).
~()IIIC C(lSCS, gave /IS t:
A report by the Defense..
tlie ~~(~~'~eBoard Task Force on Mil ita:y Soft ware ( 1987)
d field lesting for developing software: "Us-
emphasizes bollI rap' 'd
f ff l and wisdom, accurate Iy d cscn ibe the opera-
prOlolyplllg an'
'I ' Y amount 0 e "orI software
crs cnnnot. WIll an •
" .
system without . b y real op-
testing
, ts for n substnnlJ,1 . . .
tionnl rcqlllf'elllCn . ". . t and iteration on the specification, ... [I It is
. e "llionnl envlt'onmen · " '- '
orators In all op I,' develop a complex software system by build-
, Ics t , snfcsr
simp 'i ",.
and even fastest 10t actual use
" •
(
and then ad diJOg function, . enhanc-
, ., I version puttIng It In 0 u '- .,
mg a muumn .': . di to the priorities that emerge from actual
ing speed, reducing SIze, etc .. accor JOg
lise"
,- . '.1 tvoi g during informal seSSIOns . Wit . I1 .IIIdivid
IVI ua I users and
You use rapru proto yplll .'
. .' .1.. I I - 'st sites Field testing applies to beta test sites. You can
focus groups allu ut ,1 P 1,1 .1.0,. • • • • ,'" d I h
. t
collect datu tIiroug I1 III er .' views (j
and qucsuonnmrcs
•
with the users tin t iroug mea-
surcmcnts and observations of the prototype. . _,
Prototype testing requires flexible, rcpctitioll<; design ~ycles. to fit users per-
ceptions and habits with the goals of generating more consistent Illterfac~ models
and useful, enjoyable tools. Achieving these goal~ help" users own the design solu-
tion and improve their: final acceptance of the invtrumcut.

4.3 . FIVE ELEMENTS OF SUCCESSFUL DESIGN

Five interdependent elements contribute to good design: cognition, ergonomics,
utility, image, 'and ownership. These five clements define the human interface once
you have established the purpose for your product and understand the user
population. They ensure a useful and functional dialogue.
A user's view of an interface (and the product) differs from yours. Both user
and designer form conceptual models of how the instrument works. Unfortunately,
those concaptual models often differ vastly. Your goal as a designer is to create one
conceptual model of the product, or at least to accommodate both models.
Fig. 4.1 illustrates the basis from which users and designers view a product
through the five elements of design. The conceptual model of a product is built on
these five elements.
If you don't make provision for all five categories=-called product interface
levels by Robert Case (l993)-your product is deficient. Cognition involves the
~en.t~l tasks and cor:nputations involved in operating a device and relates to the
mtuiuon or expectations of the user. Ergonomics deals with the traditional con-
cerns of human factors. Utility measures ease of use and is a differentiating factor
for a prod,u~t in a competitive market. Image represents more than the styling ?f a
product; It Includes the user's perceptions of the instrument and its operation.
Ownership is the level of commitment that a user exercises to usc your product,
I
ocanneu wnn varnS
 IN'I'IIIUlAj'lt
"'III( IIIIM/\N
'"~
User's
view
Cou«:,
moe (,I
,'h~"1 l)f,f,ltH,f_t I"
"lulN
Learning COlllllllol1 1"'/1111111111111
Fit 1:1{lIlIIOIllIOII I'll,,"
Use t Jllllly 1'11111111111'
Percoptlon 111\1\110 111" 1111111
cosVCommltmont OWI\()IIIIlI" III" Vl,,"

FIG, 4.1 Five clements of 1\ luunun lntcrfuce liN viewed hy tho unor nmt rho de IIHI!'1,

These five clements intertwine to provide II bUt,liN "(H' U COIICt;j>wIIIIIlOd'l ()j' II

product. The cognitive complexity lor pcrlormlng n purtlculur tllHk I1mldln;clly III
to the utility of a device. lmngc CUll often cuhuucc OWIICI'Nhlp,whllo CI'I!.()IIonllc·I
may place buttons where a user might expect them, LhuHHlmpllf'ylng U tnsk.
Users want to understand a product and have control over (JV(.!III,\', Jf' YOII,'
instrument is difficult to use, people won't use It in splto of ill{cnpublllty, f hive you
tried to use a remote control for some consumer clcctronlcs 01' PI'OBI'UIlI u VCR'I
Most people never learn to usc the full runge of functions bccuuso of tho complex-
ity of use. Perhaps you, like me, have u VCR that continuously blinks 12:0() he-
cause you have not figured out how to set the clock, Fig, 4,2 Hhow$ u 1'01111)1 • I
control for a stereo system; it has 32 identical-looking buuons. Who iN ever t',t>iIlH ,I
,\
to remember all of them'? I'

..
I
F •
, I
I
I. I I
,
I

I
I

FIG 42 •. . 'I' II 'rite billions ure SIIIIIII, hlcmlcul In shupe,

a '. Panel from a remote control for II stereo s~s ~' . . IISlUkcs 111'0 Ircqucnt.
nd regular in layout. Consequently, operalioll is confUSing. lind II. '

v\."OI II ICU VV III I vOII'S

 THE HL~L~" L\lERF.-\CE

User's Conceptual Designer's

• •
View model view

Loarnlno Information
Fit Form

Uso Function

Perception r go Senses

CosVConunitment Q\vnership Service

Fh~.4.1 Five elements of a human interface as viewed by the user and the designer.

The e five elements intertwine to provide a basis for a conceptual model of a

PI U t, The cognitive complexity for performing a particular task ties directly 'tu
I

16 .~
~Ltr;.. -- _1=' - ~ - .. - _ T A~" Anh~nrp. nwners hiD. Whil1 e ernOnOmlCS
..
v\."ClI-I'rIC;U VVIII I va m S
FI~.4.2 Panel from a remote control for a stereo system. The buttons are small, identical in shape,
an regular in layout. Consequently, operation is confusing, and mistakes are frequent

VvUI II I~U YV I U I vdmS